SEASONAL DISTRIBUTION OF HERB A G E GROWTH LN THE SOUTH C O A S T A L REGION OF BRITISH C O L U M B I A IN RELATION TO M A N A G E M E N T OF GRAZING LIVESTOCK By Derek Edward Hunt B.Sc. (Agr.) The University of British Columbia, 1983 A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Animal Science) We accept this thesis as conforming to the required standard The UNIVERSITY OF BRITISH C O L U M B I A SEPTEMBER 1988 © Derek Edward Hunt, 1988 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of A N I M A L S C I E N C E The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date O C T O B E R 6 1988 DE-6(3/81) ii A B S T R A C T The seasonal d i s t r i b u t i o n of herbage g rowth has par t i cu la r impor tance for the management of graz ing animals since the major objective of most pasture uti l ization plans is to provide sufficient feed for continuous grazing for as great a portion of the year as possible. Providing a more even dis t r ibut ion of herbage product ion has obvious pract ical advantages. The main objective of this study was to examine the annual growth pattern of herbage in the South Coastal region of Bri t ish Columbia as affected by cul t ivars , cutting management and fert i l izer management, with the aim of extending herbage production in the fall and increasing herbage ava i l ab i l i t y dur ing the summer months. These investigations were conducted in plot trials at three different sites. Forage quality was also examined at two of these sites. In addition to these plot cutting trials an investigation into the accuracy of a height/density meter (disc meter) was conducted. The disc meter, and other s imilar non-destructive devices, have great potential as both research and farm management aids. The first plot cutting tr ial was conducted at A g a s s i z , B . C . and examined the product ivi ty of two orchardgrass (Dactylis glome rata L . ) cult ivars, Pra i r ia l and Sumas and two perennial ryegrass (Lolium perenne L . ) cu l t iva r s , Ba r l ano and N o r l e a under two cut t ing regimes ( low-infrequent and high-frequent). A n n u a l yields averaged over the three years for l ow- in f r equen t ( L I ) cu t t ing showed s ign i f i can t (P<0.05) differences between cultivars with the fo l lowing ranking: Prair ial (10,644) > Sumas (9,537) > Barlano (8,448) > Nor lea (6,666 kg D M / h a ) . Annual i i i yields averaged over two years for high-frequent (HF) cutting also showed s igni f icant (P<0.05) differences between the cul t ivars wi th the same ranking: Prair ial (9,390) > Suinas (8,625) > Barlano (7,686) > Norlea (5,953 kg D M / h a ) . The same ranking of cultivars in terms of annual yie ld was obtained in each harvest year for both L I and H F cutting treatments. A n n u a l herbage y ie lds over the three years of the tr ial showed considerable variat ion in response to c l imat ic factors. Average annual y i e ld s i n 1984 (10,116) and 1986 (10,237 kg D M / h a ) were not significantly different (P>0.05). However, yields in 1985 (5,916 kg DM/ha) were significantly (P<0.01) lower than 1984 and 1986 due to a wet, cool spring and except ional ly dry summer. Cut t ing regime d id not have a signif icant effect on annual y ie lds . There were no significant (P>0.05) differences between average annual yields under H F (7,914) and L I (8,824 kg D M / h a ) cutting, nor were there any significant (P>0.05) differences in annual yields between cult ivars due to cutting treatments. There was a tendency for L I cutt ing to produce higher yields than H F cutting for P r a i r i a l , Sumas and Bar lano wi th the reverse being true for Nor l ea . Examina t ion of product ivi ty on a seasonal basis indicated differences in yie ld due to cutting regime, depending on the season. Dur ing the spring L I cutting produced greater yields than H F cutting due to the greater yields obtained from the longer regrowth period and lower cutting height (Reid and M a c L u s k y , 1960; B l a n d , 1967; A n s l o w , 1967). However , during the dry summer months the reverse situation was observed with H F cutting producing higher yields than L I cutt ing. Sucl i a reversal in y ie ld is attributable to the dry condi t ions under which high cutt ing has been shown to produce more herbage than low cul l ing (Janti and Kramer, 1956; Appadurai and Holmes, 1964). Under L I cutting orchard grass produced 34 and 46% of annual y ie ld in the spring for 1984 and 1985 respectively and 52 and 40% of annual yie ld in the summer for the same two years. This distr ibut ion was changed under H F cutting with 21 and 36% of annual y ie ld produced in the spring for 1984 and 1985 respectively and 65 and 50% of annual y ie ld produced in the summer for the same two years. The situation was s imi lar for perennial ryegrass. Under L l cutting perennial ryegrass produced 53 and 58% of annual y ie ld in the spring for 1984 and 1985 respectively and 25 and 30% in the summer for the same two years. Distr ibut ion of annual y ie ld was more even under H F cutting with 28 and 35% of y ie ld produced in the spring of 1984 and 1985 respectively and 63 and 50% produced in the summer of the same two years. Variat ion in the dis tr ibut ion of annual production was also evident between orchardgrass and perennial ryegrass. Both orchardgrass cult ivars produced a greater por t ion of their y i e l d dur ing the summer, whereas both perennial ryegrasses produced a greater portion of y ie ld during the spring and early summer. Growth curves were developed for each cultivar for both cutting treatments over a l l harvest years which indicate the var iab i l i ty in the seasonal dis tr ibut ion of annual y ie ld attributable to c l imat ic factors and which can be varied by cutting management and choice of cultivars. Another trial conducted at the U . B . C . Research Farm #2, Oyster River, examined the effect of spli t nitrogen (N) applicat ions on annual and seasonal yields . Under regime I a total of 100 kg N/ha was applied in A p r i l . Regimes II, III and I V involved application of a total of 300 kg N / h a at vary ing times over the growing season. App l i ca t ion of 300 kg N/ha produced s ignif icant ly (P<0.05) higher annual yields (mean, 10,228) than appl ica t ion of 100 kg N / h a (7,706 kg D M / h a ) . A n n u a l yields produced under regimes II, III and I V were not s ignif icant ly (P>0.05) different (10,248, 10,245 and 10,192 kg D M / h a respect ively) . The seasonal d is t r ibut ion of y i e ld was affected by N appl ica t ion regime. Reg ime III produced a greater portion of annual y i e ld during August (20.1%) compared to the other three regimes (mean, 13.4%) and application of 100 kg N/ha in August produced significantly (P<0.05) higher yields for regime I V (1,104) compared to the other three regimes (mean, 426 kg D M / h a ) . The August N application also produced signif icant ly (p<0.05) higher yie lds in the fo l lowing spring for regime IV (2,774) than yields obtained for regimes II and III (mean, 1,810) and regime I (1,341 kg D M / h a ) . Such y ie ld improvements in the spring could have been due to improved root reserves and/or root mass produced from the August N a p p l i c a t i o n . T w o other trials conducted at Abbots fo rd also examined cutting treatments and N applicat ion regimes. The cutting trial examined the yields produced by four cutting regimes examining: high-frequent (HF) , high-infrequent (HI) , low-frequent ( L F ) and low-infrequent (LI ) cutting combinat ions . Annua l yields were s ignif icant ly (P<0.()5) lower for L F (6,721), HI (6,066) and H F (5,674) cutting regimes than the L I (8,207 kg D M / h a ) regime. L o w cutt ing (7,464) produced s igni f icant ly (P<0.05) greater yields than high cutting (5,870 kg D M / h a ) whi le there were no s ignif icant (P>0.05) differences between frequent (6,197) and infrequent cutting (7,137 kg D M / h a ) . Un l ike the Agass iz trial , high cutting produced no y i e l d advantage du r ing the dry summer months due to the exceptionally dry period where growth stopped under al l cutting regimes. The Abbots ford fer t i l izer trial examined split N appl icat ion in a similar manner to the Oyster River trial. However, due to the dry summer there was litt le response to applied N throughout most of the summer and v i thus little variation in distribution ol" annual y ie ld . Regimes II, III and IV involved application of a total of 300 kg N/ha and produced significantly (P<0.05) more herbage (mean, 5,584) than the application of 0 N under regime I (3,758 kg D M / h a ) . Annua l yields d id not differ significantly (P>0.05) between regimes II, III and I V . Results from the four cutting trials indicate that the annual and seasonal distribution of y ie ld can be affected by management factors such as variety and cul t ivar selection, fert i l ization management and defoliation m a n a g e m e n t . Forage quality was examined at both the Agass iz and Oyster River sites. Differences in forage quality were noted as a result of year, seasonal, cul t ivar , fert i l ization and cutting treatment effects. Inves t igat ions were also conducted to examine the use of a height /densi ty disc meter to measure herbage y i e l d . A s ignif icant relationship (P < 0.001) was found between herbage y ie ld and disc height for vegetative growth but more mature growth, with stems and seed heads, yielded a poor relationship. The use and accuracy of the instrument is discussed. v i i T A B L E O F C O N T E N T S Page Abstract i i List of Tables xii List of Figures xvi Acknowledgements xviii I Introduction 1 II Review of the Literature 4 2.1 Application of Growth Rate Analysis 4 2.2 Review of Techniques for measuring herbage productivity 8 2.2.1 Small Plot Cutting Experiments 10 2.2.1.1 Design I: Repeated cuttings of the same plots at time intervals of several weeks 10 2.2.1.2 Design II: Splitting of replications in series 12 2.2.1.3 Design III: Uninterrupted Growth 14 2.2.1.4 Design I V : Method of Brougham and Glenday. . . . l4 2.2.1.5 Design V : Introduction of Management treatments 16 2.2.1.6 Calculations of Growth Rates 21 2.2.2 Measurement of Pasture Growth under grazing situations 22 2.2.2.1 Direct Methods -- Grazed plots without exclosures 23 2.2.2.2 Direct Methods -- Grazed plots with exclosures (Cage Techniques) 32 2.2.2.3 Indirect Methods 47 v i i i III Experimental Materials and Methods 56 3.1 Agassiz Cutting Trial 56 3.2 Abbotsford Cull ing Trial 60 3.3 Abbotsford Fertilizer Trial 63 3.4 Oyster River Fertilizer Trial 64 3.5 Forage Quality 67 3.6 Disc Meter Investigations 68 IV Results and Discussion 70 4.1 Climatic Data 70 4.2 Agassiz Cull ing Trial 73 4.2.1 Annual Dry Matter Yie ld 73 4.2.2 Annual Dry Matter Y i e l d - cutting treatment effects 84 4.2.2.1 Cutting Frequency 84 4.2.2.2 Cutting Height 85 4.2.3 Seasonal Distribution of Y ie ld 87 4.2.3.1 Beginning of Spring Production 88 4.2.3.2 Spring Production 89 4.2.3.3 Summer Production 94 4.2.3.4 Fall Growth 96 4.2.3.5 Length of the Growing Season 96 4.3 Abbotsford Cutting Trial 102 4.3.1 Annual Yields 102 4.3.2 Seasonal Yields 102 4.4 Oyster River Fertilizer Trial 110 4.4.1 Annual Yields 110 4.4.2 Seasonal Distribution of Y ie ld 114 4.4.2.1 Spring Production 114 4.4.2.2 Summer Production 115 4.4.2.3 Fall Production 116 4.4.3 Herbage Nitrate Levels 120 4.4.4 Nitrogen Application and Regrowlh Period 121 4.4.5 Distribution of Yield and effect of Moisture 122 4.4.6 First Harvest Yields in 1986 124 4.4.7 Botanical Composition 125 4.5 Abbotsford Fertilizer Trial 128 4.5.1 Annual Yields 128 4.5.2 Seasonal Distribution of Yield 129 4.5.2.1 Spring Production 129 4.5.2.2 Summer Production. 129 4.5.2.3 Fall Production 130 4.6 Forage Quality 135 4.6.1 Crude Protein 135 4.6.1.1 Averages and Ranges 138 4.6.1.2 Agassiz 140 4.6.1.3 Oyster River 142 4.6.2 Acid Detergent Fibre (ADF) 145 4.6.2.1 Agassiz 146 4.6.2.2 Oyster River 149 4.6.3 Ash 149 4.6.3.1 Averages and Ranges 150 4.6.4 Calcium 153 4.6.4.1 Averages and Ranges 153 4.6.4.2 Agassiz 153 4.6.4.3 Oyster River 154 4.6.5 Phosphorus • 156 4.6.5.1 Averages and Ranges 156 4.6.5.2 Agassiz 157 4.6.5.3 Oyster River 159 4.6.6 Potassium. 160 4.6.6.1 Averages and Ranges 160 4.6.6.2 Agassiz 162 4.6.6.3 Oyster River 162 4.6.7 Sodium , 164 4.6.7.1 Averages and Ranges 164 4.6.7.2 Agassiz 165 4.6.7.3 Oyster River 167 4.6.8 Magnesium 168 4.6.8.1 Averages and Ranges 168 4.6.8.2 Agassiz 171 4.6.8.3 Oyster River 171 4.6.9 Copper 172 4.6.9.1 Averages and Ranges 172 4.6.9.2 Agassiz 175 4.6.9.3 Oyster River 176 4.6.10 Manganese 177 4.6.10.1 Averages and Ranges 177 4.6.10.2 Agassiz 177 4.6.10.3 Oyster River 180 4.7 Disc Meter Investigations 180 V Conclusions 186 VI Suggestions for Further Research 188 xi Bibliography 189 Appendix 1: Dry Matter Yields at each harvest, Agassiz, 1984-1986 212 Appendix II: Forage Quality Information at each harvest, Agassiz, 1984 and 1985 218 x i i L i s t of T a b l e s T a b l e Page 1 Harvest dates for high-frequent cutting treatments Agassiz, 1984 and 1985 58 2 Harvest dates for low-infrequent cutting treatments Agassiz, 1984, 1985 and 1986 , 58 3 Abbotsford fertilizer trial treatments 63 4 Oyster River fertilizer trial treatments 65 5 Annua l and three year average yields for low-infrequent cutting treatment, Agassiz, 1984, 1985 and 1986 76 6 Annua l and two year average yields for high-frequent cutting treatment, Agassiz, 1984 and 1985 77 7 A n n u a l yields under low-infrequent and high-frequent cutting treatments, Agassiz, 1984 to 1986 78 8 Botanical composition of Agassiz plots 82 9 Annual y ie ld for both cutting treatments, Agass iz 1984, 1985 and 1986 83 1 0 Comparison of Agassiz cutting treatments 84 1. 1 Seasonal production at Agassiz, 1984 to 1986 92 1 2 Y i e l d at first harvest, Agassiz 1986 93 13 Annua l y ie ld for Abbotsford cutting trial, 1985 104 14 Annua l cutting treatment yields, Abbotsford, 1985 104 1 5 Individual harvest yields for Abbotsford cutting trial 105 1 6 Y i e l d at first and last harvests, Abbotsford cutting trial, 1985 106 17 Dry matter yield for Oyster River fertilizer trial, 1985 1 12 x i i i 1 8 Percent distribution of annual yield at Oyster River , 1985 123 19 First harvest yield at Oyster River, 1986 125 2 0 Percent ground cover at Oyster River , September, 1985 127 2 1 Y i e l d for Abbotsford fertilizer trial, 1985 131 2 2 Percent distribution of y ie ld , Abbotsford fertilizer trial, 1985 134 2 3 Composition of local forages 136 2 4 Average protein content of herbage at Agass iz , 1984 and 1985 139 2 5 Average protein yield of herbage at Agass iz , 1984 and 1985 ....141 2 6 Protein content for Oyster River fertilizer trial, 1985 143 2 7 Total protein yield for Oyster River fertilizer trial, 1985 144 2 8 Average A . D . F . content of herbage at Agassiz, 1984 & 1985 146 2 9 A . D . F . content for Oyster River fertilizer trial, 1985 147 3 0 Average ash content of herbage at Agassiz , 1984 & 1985 151 3 1 Ash content for Oyster River fertilizer trial, 1985 152 3 2 Average calc ium content of herbage at Agassiz , 1984 and 1985 154 3 3 Ca lc ium content for Oyster River fertilizer trial, 1985 155 3 4 Average phosphorus content of herbage at Agass iz 1984 and 1985 157 3 5 Phosphorus content for Oyster River fertilizer trial, 1985 158 3 6 Average potassium content of herbage at Agass iz , 1984 and 1985 161 3 7 Potassium content for Oyster River fertilizer trial, 1985 163 x iv 3 8 Average sodium content of herbage at Agassiz , 1984 and 1985 165 3 9 Sodium content for Oyster River fertilizer trial, 1985 166 4 0 Average magnesium content of herbage at Agass iz , 1984 and 1985 '. 169 4 1 Magnesium content for Oyster River fertilizer trial, 1985... 170 4 2 Average copper content of herbage at Agass iz , 1984 & 1985... 173 4 3 Copper content for Oyster River fertilizer trial, 1985 174 4 4 Average manganese content of herbage at Agass iz , 1984 and 1985 178 4 5 Manganese content for Oyster River fertilizer trial, 1985 179 4 6 The relationship between the y ie ld of dry matter (lOOkg/ha) and the height of the herbage (cm), 1986 184 4 7 Y i e l d for low-infrequent cutting treatment, Agass iz , 1984 213 4 8 Y i e l d for high-frequent cutting treatment, Agass iz , 1984 214 4 9 Y i e l d for low-infrequent cutting treatment, Agass iz , 1985 215 5 0 Y i e l d for high-frequent cutting treatment, Agass iz , 1984 216 5 1 Y i e l d for low-infrequent cutting treatment, Agass iz , 1986 217 5 2 Crude protein content of Agassiz herbage, 1985 & 1985 219 5 3 Crude protein yield of Agassiz herbage, 1985 & 1985 221 5 4 A . D . F . content of Agassiz herbage, 1985 & 1985 223 5 5 Ash content of Agassiz herbage, 1985 & 1985 225 5 6 Calc ium content of Agassiz herbage, 1985 & 1985 227 5 7 Phosphorus content of Agassiz herbage, 1985 & 1985 229 5 8 Potassium content of Agassiz herbage, 1985 & 1985 231 5 9 Sodium content of Agassiz herbage, 1985 & 1985 233 XV 6 0 Magnesium content of Agassiz herbage, 1985 & 1985 .235 6 1 Copper content of Agassiz herbage, 1985 & 1985 237 6 2 Manganese content of Agassiz herbage, 1985 & 1985 239 x v i L i s t of F igu res F i g u r e Page 1. Growth Curve 7 2. Detail of Disc Meter 54 3. Disc meter recording settling height of herbage 55 4. Field plan of Agassiz Cutting Trial 59 5. Field plan of Abbotsford Trials 62 6. F ie ld plan of Oyster River Fertilizer Trial 66 7. Average Monthly Temperature, Agassiz, 1984 71 8. Average Monthly Temperature, Agassiz, 1985 71 9. Average Monthly Temperature, Agassiz, 1986 72 10. Month ly Precipitation, Agassiz, 1984, 1985 and 1986 72 11. Annual Y i e l d at Agassiz , Low-Infrequent Cu l l i ng , 1984 to 1986 79 12. Annual Y i e l d at Agassiz , High-Frequent Cutting, 1984 and 1985 80 13. Annua l Y i e l d at Agass iz , Low-Infrequent Cutting, 1984 and 1985 81 14. Growth Curve for Prairial and Sumas, Agassiz, 1984 97 15. Growth Curve for Barlano and Norlea, Agassiz, 1984 97 16. Growth Curve for Prairial and Sumas, Agassiz, 1984 98 17. Growth Curve for Barlano and Norlea, Agassiz, 1984 98 18. Growth Curve for Prairial and Sumas, Agassiz, 1985 99 19. Growth Curve for Barlano and Norlea, Agassiz, 1985 99 20. Growth Curve for Prairial and Sumas, Agassiz, 1985 100 xv i i 21. Growth Curve for Barlano and Norlea, Agassiz, 1985 100 22. Growth Curve for Prairial and Sumas, Agassiz, 1986 101 23. Growth Curve for Barlano and Norlea, Agassiz, 1986 101 24. Annual Yield at Abbotsford Cutting Trial, 1985 107 25. .Yield at each harvest, Abbotsford, Frequent Cutting, 1985 108 26. Yie ld at each harvest, Abbotsford, Infrequent Cutting, 1985 109 27. Annual Yield at Oyster River 113 28. Growth Curves for Oyster River 118 29. Yield at each harvest, Oyster River, 1985 119 30. Annual Yields for Abbotsford Fertilizer Trial 132 31. Yield at each harvest, Abbotsford Fertilizer Trial, 1985 133 32. Abbotsford Growth Curve for Fertilizer Treatments 134 33. Farm #2 Disc Measurements, June 27, 1986 182 xvi i i ACKNOWLEDGEMENTS I would like to extend my gratitude to the following people, whose help made the completion of this thesis possible: Dr . R . M . Tait - the research supervisor (Department of Animal Science), Don Bates ( B C M A & F) , Stan Freyman and Niels Holbeck ( U B C Farm #2, Oyster River). Thanks also to Silvia Hunt, Louise Janes, Petra Sanderson and Sylvia Leung for their help with sampling and lab analysis. Special thanks to Si lvia Hunt for assistance with typing and editing and H . E . Hunt for providing computer resources. The National. Sciences and Engineering Research Counci l of Canada, the University of British Columbia and the British Columbia Agricultural Sciences Council are thanked for their financial support. 1 I. I N T R O D U C T I O N In the temperate regions of the world, grassland predominates as an agricultural crop. Herbage provides the single most important source of feed for ruminants. Grazing is the most common method of harvesting and is also the most economical method of utilization. The South Coastal Region of British Columbia has a temperate climate and is well suited to the production of forage crops. The area is one of the principal dairy farming regions of Canada and also contains many small scale beef, sheep and horse enterprises. However, even with a very reasonable climate and good grassland production, large quantities of alfalfa hay (in excess of 100,000 tonnes annually) are imported from the north western U . S . A . to supply nutrient requirements primarily for dairy herds. Importation of alfalfa is seen largely as a need to improve feed quality due to the perceived lack of quality in locally grown forages. "In order to decrease this apparent need for importing high quality conserved forage, it is necessary to develop an improved understanding of the productivity and nutritive value of locally grown forage grasses in relation to their management" (Fairey, 1985a). Considerable work has been conducted locally to improve the production and quality of conserved forages (as hay and silage) through evaluation of varieties, cultivars, fertilization and other management factors. However, little work has been conducted to improve the quality, yield, utilization and management of grazed pastures. A major consideration in the use of fresh herbage from pastures for livestock feed is the pattern of production in relation to nutrient requirements. Temperate grasses have a very characteristic pattern of production. In the South Coastal Region of Brit ish Co lumbia low 2 temperatures prevent grass growth for three to four months of the year. Growth accelerates rapidly to a spring peak (usually late Apri l / early May) for several species and varieties. It is not unusual for this spring peak to occur as little as six weeks after the onset of spring growth. From the late spring maximum to late summer, production begins to slow down even though solar radiation is at its highest and temperatures approach values that are at or near the apparent optima for temperate grass growth. Lack of soil moisture and nitrogen often contribute to this 'midsummer depression'. Yet, even with adequate moisture and nutrients, there is a depression in midsummer production between the spring peak and a smaller fall peak. This midsummer depression is significantly affected by the "relative prominence of the rhythm of f lowering and tiller regeneration" (Anslow, 1965). The main objective of most pasture utilization plans is to provide sufficient feed for continuous grazing of livestock for as great a portion of the year as possible. Management practices are designed to increase the length of the grazing season and provide more feed through midsummer depression periods. The main objective of this study was to examine the annual growth pattern of herbage in the South Coastal Region of British Columbia as affected by cultivars, cutting management and fertilization management, with the aim of extending herbage production in the fall and increasing herbage availability during summer months. A n improved knowledge of annual growth patterns for various varieties and how these patterns can be altered by management factors would be very useful. It would be possible to select pasture species (monoculture or mixtures), the area seeded to each and suitable management -strategies in order to provide a more even distribution of herbage production throughout a greater portion of the year. The development of seasonal growth curves for a specific agro-ecological area would also be very useful in weekly and monthly feed budgeting of pasture and conserved forage. 4 IT. R E V I E W O F T H E L I T E R A T U R E 2.1 A P P L I C A T I O N O F G R O W T H R A T E A N A L Y S T S Several researchers working with pure swards, mixed swards and grazed pastures have generated information on the seasonal distribution of herbage yield and total annual yield. The parameters examined have varied considerably but the overall aim has been to investigate how the seasonal distribution of yield and total yield can be altered. Some of the factors that have been examined are: - varietal and cultivar differences - effect of cutting or grazing frequency - grazing management strategies - simulated grazing management strategies (simulated by cutting) - forage mixtures - influence of types of harvesting equipment - geographical location - climate / topography - soil moisture / irrigation - fertilizer application / liming Many researchers, even though they have collected data on the seasonal distribution of y ie ld , have only reported annual totals. Information on the seasonal distribution of yield can be very useful for practical farm management, and it is with this aim that other researchers have provided and discussed applications of the seasonal distribution of herbage yield. 5 For farm situations where livestock acquire a considerable portion of their feed requirements from pasture, it is desirable to have a more even distribution of pasture supply. This is especially true for midsummer and fall grazing and in mild climates winter growtli is also desired. It is hoped that by selection of appropriate varieties and management strategies, that herbage supply can be made more even throughout the grazing season. Gardner and Alburquerque (1965) measured the seasonal growth of several forage varieties in Uruguay with the aim of combining various species and grazing strategies to provide a high level of animal feed throughout the season. Anslow and Green (1967) developed seasonal growth curves for several species and varieties to compare and examine production at different times of the year, especially in midsummer. Winter and summer performance of orchardgrass (Dactylis glomcrata L.) lines was examined in Australia (Knight, 1968). Perennial ryegrass (Lolium perenne L . ) is a major component of New Zealand pastures but is susceptible to drought conditions. Summer drought conditions cause significant yield reductions in perennial ryegrass and may even lead to persistency problems. Anderson et al. (1982) generated seasonal yield distribution and total seasonal yield information for two perennial ryegrasses and two tall fescues (Festuca arundinacea Schreb.) at four regions in New Zealand to examine the potential of tall fescue for increasing summer production and sward persistency in areas where ryegrass was susceptible to drought. Cultivars of Italian ryegrass (Lolium multiflorum Lam.) give higher yields and earlier spring growth than those of perennial ryegrasses, but lack persistency. It would be desirable to combine the early growth and high yield of Italian ryegrass with the persistency of perennial ryegrasses. Ahloowalia et al. (1981) compared 6 the spring performance of 115 hybrids of Italian ryegrass cultivars. Brougham (1959) and Glenday (1959) separated growth caused by short term climatic fluctuations and smoothly changing climate and weather effects. The smooth curves enable growth rate curves to be developed for any season or the entire year, thereby allowing an estimate of seasonal and annual dry matter ( D . M . ) yield to be determined for a particular location. Morely (1968) used the growth curves developed by Brougham (1956a) to compare a number of hypothetical pastures of varying growth rates and D . M . contents under rotational grazing managements with different rest period duration and number of subdivisions. The effect of management decisions such as rest periods and number of paddocks could be varied to examine the effect on seasonal and total yield and help make management decisions to permit feed budgeting of herbage. Hal l (1973) also used the growth curves developed by Brougham (1956a) to illustrate the three phases of pasture growth (Figure 1) . It was suggested that the grazing management objective of New Zealand dairy farmers be to maintain herbage growth in Phase II through most of the season with closer defoliation occurring in the spring and fall to promote tillering. Goodenough et al. (1984) used growth curves to evaluate various seeding practices of Italian ryegrass in South Africa. Heard and Wiseman (1973) showed how annual production curves for pasture species can be used in pasture planning. With equations for grass growth curves, they used computers to facilitate calculat ions . T h e y matched stock requirements with different pasture types (i.e. species and mixtures) and examined different fields of single species or simple mixes for grazing or conservation at different times of the year. In order to conduct proper pasture planning they suggested that research be conducted to: (i) establish production curves for the principal pasture species in the main agro-ecological areas, (ii) establish the changes in quality of the forage obtained throughout the year, especially in relation to fertilizer use, management and conservation practices, and (iii) establish the optimum amounts of pasture lo be offered in balance with other feeds for maximum profit. FIGURE I: GROWTH CURVE Net G r o w t h B-B- • D n n Pha se 111 10 20 Herbage He i gh t (cm)" 30 Jones and Bartholomew (1973) applied equations to forage cumulative yields. The functions were specified with three or more data points (yields at known times). They found a good correlation between pasture experimental data and the function curves. Sucii curves can be used in pasture planning as described above. It was suggested that yields in herbage investigations should no longer be recorded strictly on an annual basis. The seasonal distribution of herbage yield gives much more information and can be readily used in pasture planning and feed budgeting. Jones and Bartholomew (1973) also suggested that old data should be searched, presented on a seasonal basis and have equations fitted to represent the seasonal distribution of yield. 8 2.2 R E V I E W O F T E C H N I Q U E S F O R M E A S U R I N G H E R B A G E P R O D U C T I V I T Y Most of the techniques used in field research on herbage productivity were developed from the 1930's to the 1960's. Many articles refer to these techniques while more recent publications do not describe methodology to any degree of detail. This is especially true for cutting techniques and herbage intake measurements. Reviews of techniques for measuring herbage productivity have been compiled by Brown (1954), Meijs (1981), Hodgson et al. (1981) and Leaver (1982). A review of numerous research articles reveals that there are several techniques for measuring herbage productivity and that the appropriate technique can differ depending on the particular parameter(s) being examined. Kemp (1973) outlined a well defined process that is used worldwide to evaluate herbage productivity for any environment. Briefly, the process can be outlined as follows: The first phase is the evaluation of herbage varieties as spaced plants. This allows a large number of varieties and/or variety combination to be assessed in a relatively small area. This procedure is used mostly to assess the broad adaptation of a number of varieties to the environment and is used almost exclusively for initial variety screening. The next stage is the evaluation of small swards in plots. Varieties are evaluated under a wide range of conditions. Treatment effects such as fertilization, irrigation, seeding rates and simulated grazing using cutting techniques are also commonly evaluated using small plot experiments. Only occasionally are animals introduced to examine their effect on treatments or test species. It is with small plot experiments that most seasonal growth curves have been developed using cutting techniques. 9 The third stage is the determination of animal performance from test species, fertilization regimen, cutting trials, etc. Most of these trials involved testing of different species or species combinations and many involved the cutting of herbage and subsequent feeding to livestock. However, in several trials the herbage was harvested directly using grazing livestock on herbage plots. This third stage of the evaluation program is the main test. It is the most important since the ultimate objective of improv ing forage performance is to improve animal performance. The final stage of the program is to test and evaluate under commercial conditions. Whether this is clone on research farms or commercial farms, the aim is to evaluate under practical conditions that dominate commercial farming. The fol lowing review of the literature examines studies in the following categories: 1. Smal l Plot Cutting experiments using small herbage plots to evaluate varieties, climate and numerous management factors. 2. Large plot experiments that have grazing animals introduced to evaluate herbage productivity under grazing conditions for different varieties, climate effects and numerous management factors. 3. Indirect or Non-Destructive Methods. A l l of the selected papers have presented information on seasonal growth of species used for intensive animal production. The experiments were either specifically designed to determine seasonal growth of pasture species or the information was gained as a secondary objective to the examination of different management treatments such as irrigation, fertilizer regimen, cutting regimen, grazing regimen, etc. 2.2.1 Small Plot Cutting Experiments The small plot trials examined produced seasonal growth information as a result of harvesting over time. A l l these trials are categorized as 'Direct Methods' since they involve actual herbage harvest. Kemp (1973) organized the experimental design into five categories described as follows: Design I: Repeated cuttings of the same plots at time intervals of several weeks. Design II: Splitting of replications into series. The method of Anslow and Green (1967). Design III: Uninterrupted growth. Design IV: Method of Brougham (1955, 1956a, 1959) and Glenday (1955, 1959). Uninterrupted growth in time and space replicates. Design V : Introduction of management techniques. These designs are described in more detail in the following pages. 2.2.1.1 Design I: Repeated cuttings of the same plots at time intervals of several weeks. This is the simplest and most commonly used plot harvesting technique. Repeated cuttings of the same plots are conducted at time intervals of several weeks. A l l plots are cut at the same time. The information produced can be used to measure long term changes in seasonal growth rates. In addition, general information is obtained on the difference in production between species and management treatments within seasons and within species over time. The main advantage of Design I is the minimal requirements for land and labour compared to other designs. Knight (1968) cut plots at monthly intervals during the summer and at two month intervals in the winter to examine growth of several orchardgrass varieties and to identify materials with good winter growth (no winter dormancy) and which were early heading. Ahloowalia et al. (1981) examined the performance of 115 perennial ryegrass / Italian ryegrass hybrids in comparison to Italian ryegrass cultivars. They were interested in spring growth, annual yield and persistency of the hybrids. Four harvests were taken each season at specified calendar dates. Kean (1982) conducted experiments to determine the seasonal growtli pattern and total seasonal yield of several grass varieties. Depending on the experiments, plots were cut at 4, 5 or 6 week intervals. Fairey (1985a) harvested plots four times each season on specified dates to evaluate the yield and quality characterization of perennial ryegrass, hybrid ryegrass (Lolium pcrenne L . x Lolium italicum A . B r . ) , orchardgrass and reed canarygrass (Phalaris arundinacea L . ) at the Agriculture Canada Research Station, Agassiz, B . C . . Method I has also been used to evaluate fertilization regimes. Castle et al. (1965) cut plots at fixed 4 to 6 week intervals to examine the effect on total yields, distribution of yield over the season, clover content and protein production of six different time schedules of applying a constant weight of fertilizer. Reid and Castle (1965) cut at four to five week intervals to examine the effects of fertilizer N , irrigation and irrigation/N interaction on herbage yield, in grass and grass/clover swards. Cowling and Lockyer (1965) cut at regular intervals to study the response of a range of grasses to N applications and their growth in association with white clover. Wolfe and Crofts (1969) determined the seasonal growth pattern of several grasses under coastal conditions in N S W , Australia by harvesting six times per year at fixed dates. They measured the effect on seasonal production of N used by each grass at different times of the year. Cameron and McGowan (1969) studied the immediate and residual effects of individual applications of superphosphate on seasonal pasture growth up to 4 years after application. 2.2.1.2. Design II: Splitting of Replications into Series. In order to provide more estimates of sward growth rates over time, Anslow and Green (1967) modified design I by using replications in time as well as replications in space. They had a series of plots which were harvested in a different sequence to produce overlapping growth periods from which they developed mean growth curves. The technique gives a more continuous record of the rate of production than Method I. Anslow and Green (1967) used 3 series of plots each with two replicates. They describe their method as follows: "To obtain frequent estimates of the rate of herbage production, only two of the six blocks were cut on any one date. Another 2 were cut after the lapse of one-third of the current growth interval and the remaining 2 after two-thirds. The procedure was then repeated on the same plots and in this way a rotational cutting pattern was established with three pairs of plots being cut in staggered sequence and remaining permanently out of phase. Every 'x' days an estimate could thus be obtained of the amount of production which had accumulated in '3x' days" (Anslow and Green, 1967). The method was used to determine the distribution of yield for a range of grasses. In early May cutting intervals of 21 days were used to prevent inflorescence emergence in perennial ryegrass and tall fescue. The interval was extended to 28 days as the growing season progressed and in late fall and early winter longer cutting intervals were used. Anslow (1967) used the technique to compile production curves for a number of species and varieties with particular reference to midsummer growth with no limitations of water or fertilizer. Morris and Thomas (1972) made a slight modification with a 6 week cutting interval per series (3 separate scries) to examine seasonal patterns of production of widely different grass types over a range of environments. Corrall and Fenlon (1978) also modified the method so that the cutting interval was always one week between series, to allow for easier cutting and fertilizer management. However, the 4 week interval did not always prevent stem development. Corra l l and Fenlon (1978) generated seasonal distribution curves of herbage production for various varieties and management factors. Corrall (1978) examined the pattern of production of different grass species and the effects of irrigated versus non-irrigated swards. Orr et al. (1988) compared production curves generated by method II with patterns produced under continuous grazing by sheep and noted that the curves produced by these two methods were fundamentally different. It was suggested that it is not appropriate to relate production curves generated by a cutting regime to the potential under continuous grazing. They also added that in practice a more uniform seasonal pattern of production is seen in swards maintained on the basis of controlled sward surface heights and that it is easier to match feed supply willi feed requirements as the season progresses. 1 4 2.2.1.3. Design III: Uninterrupted Growth. The growth rales of pasture species has also been measured as the uninterrupted growth of the species where the same area is not harvested twice. Featherstone et al. (1951) mowed off an area in a pasture and then harvested different plots twice a week for twelve weeks. Growth was measured on a mixed sward to examine variation in composition as growth proceeded from early A p r i l until the end of June. The aim was to determine when the best yield per hectare could be obtained from a single conservation cut. Henzell and Stirk (1963) cut quadrats in plots at regular intervals to measure the relative effects of soil moisture stress and nitrogen deficiency on the growth of sown pasture in south east Queens land. Norman (1963) examined uninterrupted growth by harvesting quadrats at 2 week intervals to examine changes in dry matter yie ld and nitrogen and phosphorus content of native pastures in the Northern Territories of Australia. Measurements were made during the summer grazing season and the early part of the subsequent dry season. Calder and Davies (1965) measured uninterrupted growth every 2 weeks in order to investigate the onset of spring growth of early and late types of perennial ryegrass and orchardgrass. 2.2.1.4. Design IV: Method of Brougham and Glenday. Brougham (1955, 1956a, 1959) and Glenday (1955, 1959) modified Design III to include replication over time as well as replication in space. A series of plots is mown off and uninterrupted growth is measured on the plots over a period of several weeks. No one plot is harvested more than once. Every one to two weeks a new set of plots is commenced so that growth curves for the same stage of herbage growth are produced over a period of time. The method has not been widely adopted and was used mostly by Brougham and Glenday for detailed analysis of herbage growth rates. Brougham (1955) conducted weekly cuts to determine the nature of the growth curve of a ryegrass-clover pasture and to obtain information on the cutting technique, its statistical design and analysis. The technique was used on a short rotation ryegrass and clover pasture under various resting periods during late autumn and winter to determine the influence of time of commencement and duration of resting on subsequent growth rates (Brougham, 1956b). The method was also used by Brougham (1959) to develop a set of detailed annual growth curves for total herbage (ryegrass and white clover (Tr if olium repens L.)) with the influence of short term weather fluctuations removed. Wil l iams and Biddiscombe (1964) examined winter growth of selected species in New South Wales, Australia, on north and south facing slopes of a valley to observe various temperature differences. Wilman (1965) harvested undisturbed growth to analyze the response of Italian ryegrass to different levels of applied nitrogen and to examine crop development week by week. Brougham and Glenday (1969) examined the effects of short term changes in some weather factors on daily growth rates of pure swards of perennial ryegrass, short rotation ryegrass and orchardgrass in the early regrowth stages. 2.2.1.5. Design V : Introduction of Management Treatments. The seasonal growth of pastures has also been assessed by trials that have investigated the effects of cutting on herbage yie ld . These experiments have used one of the four previously described designs and imposed further cutting treatments on the design. They have used: (a) various cutting frequencies. (b) various cutting heights. (c) cutting according to growth stage or sward height. (d) various combinations of the above three. (a) Cutting Frequency. Most experiments have determined the seasonal growth of pasture species by repeated harvests,, making it necessary to choose appropriate cutting intervals. In those experiments using Designs III and IV, the varied cutting intervals were part of the design and no choices had to be made. However, experiments using Designs I and II required decisions to be made as to the duration of the intervals between harvests. In general the interval between harvests has been based on two criteria: (i) A fixed time scale such as calender date or fixed number of days or weeks between harvests. The selection of harvest dates is not related to any stage of plant development. (ii) Cutting intervals have also been based on some stage of plant or sward development. Swards have been harvested at stages such as when the pasture would normally be grazed. Harvests are also made at the hay or silage stages. Several trials have based harvest dates on stages of plant development such as ear emergence and anthesis. Many of the trials investigating cutting interval used Design I and harvested plots at varying intervals. Oyenuga (1960) investigated the length of the recovery period (3, 6, 8, or 12 weeks) in Guinea Grass to examine effects on yield and quality characteristics. Shaw et al. (1965) examined the seasonal growth of 17 introductions of Paspalum species into south east Queensland as affected by three cutting intervals (4, 8 and 12 weeks) to impose different degrees of stress. Bland (1967) imposed three cutting treatments (2, 4 and 6 harvests per year) to examine the yield of perennial ryegrass and the clover contribution to nitrogen yield with clover and grasses planted in rows with segregated or unsegregated roots. Lambert (1962) examined the growth in swards of timothy (Phicum pratense L.)and meadow fescue (Festuca elatior L . ) under two levels of nitrogen fertilization and two cutting treatments. One set of plots was cut regularly every 4 weeks whereas the other set was harvested at the 'hay' stage and the aftermath harvested when anthesis started. Langer (1959) determined the response of Timothy and Meadow fescue to three cutting intervals by harvesting plots: every four weeks, at the 'hay' stage and at the 'silage' stage. Davis (1961) investigated the relation between maturity class and other characteristics of 40 orchardgrass varieties at Agassiz, B . C . , under a silage / pasture type of utilization. The first harvest of each year was made when maturity was assessed (the number of days from March 1st. required by each variety to subtend one-third of its inflorescence free of the sheath). Subsequent cuts were made when the forage reached a height of 30.5 cm to 35.5 cm. Gardner and Alburquerque (1965) measured seasonal growth of several forage species in Uruguay with the aim of combining various species and methods to provide a high level of animal feed throughout the year. Harvests were made when sward heights had reached 20 cm. Fairbourn (1983) determined the effects of phenological development at first harvest on the total dry matter y i e l d , of eight High Plains grass species under irrigation. First harvests were made on the basis of growth stage ( three leaf blades per tiller, anlhesis and boot heading). Subsequent harvests were made when the herbage reached heights of 25 cm to 32 cm. Reid (1985) harvested swards at a height of 15 cm to 20 cm or before ear emergence to obtain yield information on some commonly used grasses in the U . K over a wide range of nitrogen application rates. In general it has been found with both temperate and tropical grasses that infrequent cutting increases yield, but depresses quality. The optimum cutting interval for any pasture species for either quality or yield seems to be related to the stage of development of that species rather than any fixed harvesting interval. These results suggest that it is probably better to adjust cutting intervals throughout the year in accordance with grass growth. (b) Cutt ing Heights. A l l of the trials reviewed in this section defoliated the swards by cutting with a variety of motorized or hand operated equipment. Animals were sometimes introduced after sampling to defoliate the remainder of the plots and add a biotic factor (Edye, 1967). The height to which the swards were cut depended on the objectives of the authors. Langer (1959) harvested to ground level when studying sward structure and dynamics. Castle et al. (1965) harvested to grazing height in order to relate their results to grazing situations. Appadurai and Holmes (1964) examined the interaction between closeness of cutting and three soil moisture regimes (wet, medium and dry). The 'pasture' plots were harvested at a height of 20.3 cm and cut to 2.5 cm (close cutting) or 6.4 cm to 7.6 cm (lax cutting). 'Silage' plots were harvested at a height of 35.6 cm to 2.5 cm (close cutting) or 6.4 cm to 7.6 cm (lax cutting). Bryan and Sharpe (1965) were interested in maximizing forage yields and accordingly varied their cutting heights and cutting intervals. Davis (1960), again working at Agassiz, examined the effect of clipping to various heights on the yield, protein and lignin contents of reed canarygrass. The cutting treatments were designed to look at the potential of reed canarygrass under a silage or grazing management system. Calder and Davies (1965) compared orchardgrass and perennial ryegrass for early spring production. When all cultivars were harvested at 5 cm, some were classified as early and others as late. When all cultivars were harvested to ground level there were no differences in classification. Some researchers have harvested to ground level for this reason. However, ground level cutting is not the level to which animals normally graze. Studies have also been conducted to investigate the effect of cutting height and cutting machinery on yield. Black and Alexander (1967) found that at short cutting heights (3.8 cm) there was a marked depression of regrowth of all grasses studied (two perennial ryegrass and two orchardgrass varieties) after cutting with a flail harvester as compared to a reciprocating blade mower. 20 (c) Combination of Cutting Frequency and Cutting Height. Several studies have investigated combinations of both cutting frequency and cutting height. Most have used Design I with various cutting regimes. Burger et al. (1958) examined the influence of different heights and frequencies of cutting on relative yields of smooth bromegrass (Bromus incrmls Leyss.) and tall fescue when grown alone or in combination with cither ladino clover or Alfalfa (Medicago saliva L.) or both. Plots were harvested five times per year (pasture), four times (silage) and three times (hay) with three cutting heights ( to 2.5 cm, 5.0 cm and 10.0 cm ) for each cutting frequency. Reid (1959) studied the effects of two cutting heights (2.5 cm and 5.0 cm tb 6.4 cm) under two different frequencies of cutting (five or six cuts for pasture and three or four cuts for silage) with six ferti l izer nitrogen treatments superimposed. Frame (1965, 1966) compared productivity of a grass / clover sward under different regimes of cutting and grazing. Swards were defoliated by cutting or grazing to two heights ( 2 . 5 cm to 3.8 cm as the 'low' and 5.0 cm to 6.4 cm as the 'high'). Grazing or cutting was conducted monthly or when herbage growth had reached 18 cm to 23 cm. Bryan and Sharpe (1965) wanted to determine the growth pattern of pangola grass under local conditions in eastern coastal Queensland. They measured the effects of varying cutting intervals (4, 8 or 12 weeks), cutting height (2.5 cm to 5.0 cm and 13 cm to 15 cm) and nitrogen fertilizer level. Boswell and Mosgiel (1977) conducted a wide range of defoliation regimes in order to identify the most productive simulated rotational grazing management system. They tested frequent, infrequent and strategic (combination of frequent and infrequent cuttings) treatments. The strategic treatment took into account the modifying effects of season on pasture growth. Two terms of reference were used to determine frequency of cutting: when pasture reached a predetermined height or a fixed period of regrowth between cuts (two or four weeks). Two cutting severities were also compared (cutting to 3 cm or 6 cm). Anderson et al. (1982) measured the seasonal distribution of production and total seasonal production of two perennial ryegrasses and two tall fescues at four sites in New Zealand under eight treatments examining combinations of cutting height and cutting frequency. 2.2.1.6. C A L C U L A T I O N O F G R O W T H R A T E S : The changes in growth rate over the season are usually presented as weight of dry matter per unit area per unit time (i.e. kg D.M./ha/day). The methods used to determine the growth rate of a pasture have depended on how the herbage growth curve has been interpreted. Most have assumed linear growth between harvests (Williams and Biddiscombe, 1964; Langer, 1958; Anslow, 1965 and 1967; Bryan and Sharpe, 1965; Morris and Thomas, 1972; Corrall and Fenlon, 1978). The assumption of linear growth is not always valid. However, there is no alternative when using Designs I and II unless the form of the growth curve is already known from some other data. For Designs III and IV it is possible to determine the form of the growth curve. Brougham and Glenday (1967) applied logistic functions to the data when the growth pattern was sigmoid. However, it is not always possible to know the appropriate mathematical functions for every set of data. Anslow and Back (1967) applied linear functions while Brougham and Glenday (1967) applied logistic functions to different portions of related data. More recently the approach has been to fit equations by the 22 least squares method to each set of data allowing growth rale to be calculated from the fitted equation (DuChateau et al. 1972; Radford, 1967; Heard and Wiseman, 1973). In using mean growth rates as a data point between two harvest dates a major problem arises in the correct location of the point unless the shape of the curve is known. The usual and best approach has been to place the mean growth rate at the mid-point of the harvest interval. Anslow and Green (1967) developed Design II in an attempt to better define changes in herbage growth rales with time. Still, linear growth was assumed between harvests and since intervals can be several weeks, the improvement in accuracy may not be that great (Kemp, 1973). It is generally recommended that non-destructive measurements be taken between harvests, especially when harvests are infrequent. Such non-destructive measurements can be made with electronic capacitance meters, height/density meters or visual assessments. For investigations that require accurate information on short term changes in growth rate with time, Design IV is the most appropriate. 2 . 2 . 2 . M E A S U R E M E N T O F P A S T U R E G R O W T H U N D E R G R A Z I N G S I T U A T I O N S . Similar to the small plot trials discussed in the previous Section A , there have been a large number of trials that have examined varietal, cultivar and management factors under a range of grazing conditions. Not only have these experiments provided information on the seasonal distribution of yield and total yield but additional information on grazing 23 behaviour, animal productivity, and animal/sward interactions have also been obtained. Techniques that involve the actual harvest of herbage are referred to as 'direct methods' and can be categorized on the basis of whether or not exclosures were used to exclude grazing animals from caged areas. 1) Direct Methods - Grazed Plots Without Exclosures. a) Measurement of Growth during the resting period. b) Alternate Grazing and Mowing Technique. c) Sears' Technique - Return of feces in proportion to treatment product ion . d) Difference Techniques. 2) Direct Methods - Grazed Plots With Exclosures (commonly called the Cage Techniques) a) Permanent Cages b) Movable Cages -- the Pre-Trim Technique c) Movable Cages — No Trimming Cut 2.2.2.1. Direct Methods - Grazed Plots Without Exclosures. (a) Measurement of Growth During the Resting Period: Davies and Trumble (1934) and Brown (1954) described a method for measuring pasture growth on grazed plots without the use of exclosures. Animals are used simply to remove leaves, to provide the biotic factor and to maintain the plots in pasture condition. Procedure : (i) immediately prior to the animals being placed on the experimental plots of pasture land, randomly located quadrats or strips of herbage are cut. (ii) the animals are then put on at such a heavy stocking rate that the herbage is eaten down to the level of the cut areas within one or two days . (iii) once the herbage has been eaten down to the level of the sampled areas the pasture is rested for a period and the cycle repeated. Davies and Trumble (1934) described how the technique was used in South Australia by Richardson (1932) who trimmed quadrats to a uniform height both after grazing and before the commencement of the following grazing period. Dickinson et al. (1981) used a grazing period of 7 clays and estimated herbage growth during grazing using the method of Campbell (1966) in which growth rate during grazing is assumed to be the same as the rate of regrowth of the sward during the preceding rest interval. C a l c u l a t i o n s : The growth during the rest period is calculated as a daily average. The herbage sampled from the quadrats or strips is weighed and divided by the number of days of the rest period to give an average daily growth rate. The seasonal distribution of growth can therefore be plotted as a sequence of growth rates from the rest periods. The sum of the yields for all the periods is the yield for the season. The weight of the samples cut just prior to the grazing period estimates the amount of herbage available to the livestock for that grazing period. The amount of herbage grazed is not used in the calculations. When results are used for only comparative purposes the growth during the short grazing period can be ignored. However, for absolute results growth during the grazing period must be inc luded. 2 5 The rest period: The rest period will vary in length depending on the season and climate. For example, in the dry climate of South Australia, during the winter, growth rate is determined from grazing periods with rest periods of about 6 weeks, whereas under more temperate conditions in the U . K . the rest period may be 3 to 4 weeks and during periods of most rapid growth this may be as short as 2 weeks. A d v a n t a g e s : Compared to most other agronomic methods of determining pasture growth rate, this method is inexpensive and simple. Disadvantages: Lynch (1947) made certain objections to the procedure. These are as follow: (i) results are only approximate if growth during grazing is left unaccounted. (ii) it is difficult to control the livestock in order to graze exactly to the sampled height. (iii) grazing above or below the sampled height will lead to higher or lower estimates of production. Grazing leaves the field with an uneven height of pasture that is not equivalent to a mown height. (iv) a problem can arise on smaller experimental pastures of rccutting quadrats or strips. Areas of use: The method is useful for comparing the productivity of different herbage species and varieties and of different fertilizer treatments. It has been used to obtain a rough idea of herbage production under grazing and has been used to supplement stock grazing experiments. 26 fb) Alternate Grazing and mowing technique: This method was developed by Hudson (1933) in New Zealand and has also been described by Lynch (1947) and Brown (1954). The method measures pasture production continuously over the grazing season while livestock are grazing. Hudson (1933) discussed the reasons for developing the technique. Background to the technique: In many pasture investigations in New Zealand during the late 1920's the lawn mower was used. It was pointed out that 'mowing only' had profound effects on the botanical composition of the sward. It was suggested that if stock could be used together with mowers that these changes could be minimized. The continual use of a mower results in more prostrate species becoming dominant since more upright species, such as grasses, are defoliated more severely than the prostrate species such as clover and flat weeds. In contrast, under grazing, the prostrate species might be defoliated as severely or more severely than upright growing species. Mowing also fails to distinguish between the palatability of species. Under mowing a more palatable species may contribute a greater proportion of the sward than it would under grazing conditions where that species is preferentially selected by the grazing animal . Procedure: The technique involves two duplicate sections (A and B) in each of which the treatments are arranged in a series of randomized blocks (or latin squares). Two types of mowing are used: (a) ' M and W , when herbage is mown and weighed; and (b) 'm and c', when herbage is 27 mown to clear up growth to an even mower height before leaving to grow for an ' M and W cut. C a l c u l a t i o n s : The alternate mowing and grazing technique allows continuous production records to be maintained. The growth periods between cuts may range from 10 days to 2 or 3 months depending on the season. Growth rates will be averaged over these periods by dividing the total herbage accumulated for a period by the number of days in that period. From this the seasonal distribution of growth rates can be plotted and the total yield for the season obtained by summing the production for each growth period. T ime between clippings and grazing p e r i o d : As mentioned, the time between clippings will vary according to season. Also , time between clippings and grazings will depend on the growth rates. Each grazing may be in reality more or less than a day depending on the amount of herbage to be consumed and the number of animals used. The number of grazings also varies according to the pasture growth rates and weather. During wet weather the animals should be grazed more frequently to utilize herbage at a shorter stage to prevent fouling. Hudson (1933) used from 20 to 40 sheep per plot which gave a stocking rate of 200 to 400 sheep per hectare. A d v a n t a g e s : The alternate grazing and mowing technique was adopted by the New Zealand Department of Agriculture in the 1930's. It was felt that the technique was ideally suited to the investigation of certain fundamental problems. The main advantage of the method is that the effects of the grazing animal are present on the experimental area. 28 Nutrients are returned, via animal feces, to the sward and the selective effect of the grazing animal is introduced to maintain a typical sward under grazing conditions. Lynch (1947) stated that a high degree of accuracy of measurement is possible with the technique. He listed ten trials that used the technique over a total of 27 years showing the very satisfactory low standard errors associated with the results. The technique measures the maximum production of a sward under the prevailing climatic conditions, because it maintains the sward in a leafy stage below 10 to 15 cm (Hudson, 1933). Disadvantages: According to Lynch (1947) the mowing and grazing technique suffers from two serious defects: (i) The form of management allows for no growth beyond a leafy 10 to 15 cm and while this is ideal for the highly productive sward it is an impossible realization for the average farmer. (ii) There is a transference of nutrients. As the animals graze all treatments within the one enclosure and their droppings are distributed more or less uniformly over all, it is to be expected that nutrients will be transferred by this means from the highly productive plots to those of lower yield. L y n c h (1947) discussed some of the trials that were undertaken to measure this transfer of nutrients. The material showed that the effect of transference of nutrients is to reduce the measurable response from treatments to a real extent, although not to the extent thought possible by other researchers at the time. Hudson (1933) realized this problem and developed an alternative method to eliminate the transfer of fertility. 29 Areas of use: As mentioned earlier, the alternate mowing and grazing technique was felt to be ideally suited to the investigation of certain fundamental problems. The technique was used by the New Zealand Department of Agriculture in the following investigations (Hudson, 1933): (i) the effect of time of application of N and P fertilizers on seasonal product ion . (ii) a comparison of infrequent-heavy dressings versus frequent-light dressings of lime and phosphate. (iii) a comparison of the effect of different phosphate and nitrogen carriers . (iv) a comparison of newer concentrated compound fertilizers with simple mixtures. (v) determination of the relative yielding capacities of different ryegrass, orchardgrass and clover varieties as constituents of pastures. (vi) examining various combinations of lime, P and K (Lynch, 1947). (vii) Brown (1954) mentions that the technique was used in Scandinavia with dairy cattle, but with alternation of grazing and cutting occurring each year. The experiments extended over 4 to 6 year periods. (viii) more recently Baars (1976a) used the alternate mowing and grazing method to measure the seasonal distribution of pasture production over a 16 year period at Hamilton, New Zealand. Two series of plots with five time replicates at 3 to 4 day intervals were used so that alternately the plots were cut and grazed by sheep. Yields for the three to four day intervals were added to give yields for standard 14 clay cutting intervals. During periods of little growth due to drought, yields were averaged for the 28 days preceding a cut. 30 (ix) Ell iott and Lynch (1958) felt that the technique was quite satisfactory for small plot trials where the primary aim is an estimation of fertilizer response by pastures and where an accurate measure of total production is not required. (c) Sears' Technique: This technique was developed by Sears (1944) at Palmerston North in New Zealand, with the main objective being to eliminate the transference of fertility that occurs with Hudson's alternate mowing and grazing technique. Sears' technique is described by Sears (1944), Lynch (1947) and Brown (1954). The technique involves the return of herbage clippings and animal droppings in proportion to herbage growth during the resting period. Procedure: (i) all treatment plots exist in one common enclosure. (ii) when a plot is ready for grazing, a strip of herbage is mown from the plot and this cut represents the amount of herbage available for grazing. (iii) the herbage taken from the strip is weighed and then returned to the strip it was mown from. (iv) a consideration of the amount of herbage available for grazing and the size of the plot is made and an adequate number of livestock are placed on the plot so that the herbage is grazed down to the level of the mown strip in 2 to 3 days. (v) all of the animals are harnessed so as to permit urine and feces to be collected (Sears and Goodall, 1942). (vi) the feces and urine are collected, sieved, weighed and diluted with water. (vii) after grazing, the diluted feces and urine are watered back onto the grazed plots and boundary areas in proportion to the amount of dry matter produced by the herbage grown under each treatment. If any one treatment has a number of replicates, the diluted feces and urine are watered back equally on each of the replicates even if there was any variation in yield between the replicates. Determination of herbage dry matter was made while the feces and urine were being collected during the grazing. C a l c u l a t i o n s : Growth rates can be measured between grazings. An average growth rate can be calculated for each rest period by dividing the total herbage accumulated for a period by the number of days in that period. Continuous production records may be obtained as long as the treatments are adequately replicated and the common enclosures are paired as in the alternate mowing and grazing technique. Thus, both the seasonal distribution of growth rates and the total seasonal production can be obtained. A d v a n t a g e s : (i) The technique overcomes the problem of the transference of fertility encountered with the alternate grazing and mowing technique. (ii) Continuous production records can be obtained. (iii) The selective effect of the grazing animal is introduced to maintain a typical sward under grazing conditions. Nutrients are returned back to the sward.. D i s a d v a n t a g e s : The technique is labourious, time consuming and expensive. It cannot be conducted unless there are research stations with a large staff and enough money available to use the method. 2 . 2 . 2 . 2 . D i r e c t M e t h o d s - G r a z e d P l o t s W i t h E x c l o s u r e s ( C a g e T c c h n i (i u c s ) Cage techniques have been described by many researchers, including: Brown (1954), Green (1959), Grassland Research Institute (1961), Carter (1962), Cooke (1969), 't Mannetje (1978) and Boswell (1980). The cage techniques measure herbage growth within exclosures. These exclosures often take the form of cages which are placed on the pasture in one position for the entire grazing season (permanent cages) or they arc moved around the pasture to several positions (movable cages). Exclosures are used when the animals are grazing on the pasture for such a long period that the amount of herbage growth during the grazing period(s) cannot be ignored. Sample areas are protected from grazing through the use of cages. Cage techniques arc commonly conducted in conjunction with animal production trials. C a r r y i n g out the two types of experiments s imultaneously generates comparative information on both herbage production and herbage consumption. Brown (1954) advised that cage techniques should be conducted on two or three areas in rotation so that the pasture can be rested. The length of the rest and grazing periods will depend on the amount of herbage present and the growth rate of the herbage. Both seasonal and annual yields can be determined irrespective of the cage method used. (a) Permanent Cages: This is the simplest of the cage techniques. Cages are placed on the pasture and remain there until the grazing has been completed for the season. Procedure: (i) Cages are distributed throughout the pasture. (ii) The areas protected by the cages are clipped throughout the growing season. Frequency of clipping depends on the growth rate of the pasture. The number of clippings over the growing season is commonly five or six although there may be as many as ten or eleven. (iii) Clipped samples are dried and weighed to determine yield. Advantages : The method is simple and inexpensive when compared to other cage techniques. D i s a d v a n t a g e s : The method suffers from all the disadvantages of continuous clippings and lacks the animal management factor of selective grazing and the return of nutrients via urine and feces to the pasture. An eleven year experiment conducted by Brown and Munsell (1945) and a five year experiment conducted by Robinson et al. (1937) showed that permanently caged areas definitely deteriorate, with a decrease in yield and a replacement of grass by weeds which supply less dry matter. Hudson (1933) showed that an area of pasture not subjected to grazing, but rather clipped, had an increase in flat weeds and a reduction in vigour when compared to a similar grazed area. 34 Application of the T e c h n i q u e : Robinson et al. (1937) compared the responses of permanent pastures to fertilization as measured by: (a) grazing with dairy cattle, (b) clipping permanently caged plots of pasture, and (c) clipping by the difference method. The yields obtained by clipping permanently caged sites showed a progressive decrease in comparison to the yields obtained by grazing. However, there was a high correlation in any one year between the yield of permanently caged plots and the yields from grazing. Brown and Munsel l (1945) examined the chemical and botanical characteristics and yields of herbage cl ipped from caged areas in differently fertilized grazed pastures. Two cage methods were used: in one method a series of cages was placed on the pasture and moved every spring; the other method involved a series of permanently positioned cages. In each of three differently fertilized pastures over eleven years the permanently caged areas had less grass, many more weeds, more bare ground and consistently smaller yields than areas caged for only one year. Brelin (1979) used permanent cages to compare pasture yields under two different grazing management schemes of mixed grazing with cattle and sheep and single grazing. (b) Movable Cages -- The Prc-Trim Technique: With some pastures it has been found to be advantageous to cut the grazed herbage to a uniform height (a trimming cut), before placing a cage over the area and allowing it to grow again for a sampling cut. In other pastures it is found that the trimming cut is not desirable because it is too severe a shock for some plants. Brown (1954) described techniques for both lush, rapidly growing herbage and for sparse, slow growing herbage. Pre -Tr im Technique for rapidly growing herbage: This technique was practiced at the Illinois Agricultural Research Station and at several other stations in the U.S . . It has been described by Fuelleman and Burlison (1939, 1940), Nevens (1941, 1943 and 1945) and Brown (1954). P r o c e d u r e : (i) Just before the animals are placed on the pasture, an area the size of the cage is trimmed to a certain height. This weight of herbage, 'A' , is the quantity of herbage available to the animals when they are first turned onto the pasture. (ii) The cage is then placed on the trimmed area to protect it from grazing. (iii) After a certain amount of l ime, the herbage that has accumulated as regrowth on the caged, protected area is cut down to the trim height and weighed. The amount of time that the cage is left at one location from trimming to sampling can be quite variable. The regrowth can be sampled at the end of a grazing period when the animals are removed from a pasture. The length of the grazing period can vary according to the amount of herbage available for grazing on the pasture or it may be of some fixed duration. The regrowth can also be sampled while the animals are still grazing. Sampling can be conducted at regular intervals, such as every two weeks or once a month. Regrowth can also be sampled at certain phenological stages. 3 6 (iv) It must be remembered that if the cages are moved while the animals are grazing that the trimming height must be lower than the height of the grazed pasture. (v) The weight of the regrowth herbage, 'B', sampled from caged areas represents the growth during the period that the cage was on that particular site. (vi) After sampling, the cage is moved to another position that has been previously trimmed. (vii) The regrowth, ' C , sampled from this second site represents the growth for the period that the cage was on the second site. (vii) After each sampling of the regrowth, the cages are moved to a new site. This process continues for the rest of the growing season. C a l c u l a t i o n s : Growth rates can be calculated for each period that a cage was on a particular site by dividing the amount of sampled herbage by the number of days from trimming to sampling. Therefore , the seasonal distribution of growth can be obtained over the growing season. The total seasonal production can be obtained by summing all the samples. For example: Total Seasonal Production = A + B + C + .... etc. A d van tages: (i) The technique is well suited to measuring pasture yield under rotational grazing (Nevens, 1945). (ii) The technique has lower standard errors and lower coefficients of variability than the Difference Technique (Nevens, 1945). (iii) Less samples are required compared to the Difference Technique to give a standard error of 10 % of the mean (Radcliffe, 1971). D i s a d v a n t a g e s : (i) The technique is not suited for grazing experiments where herbage intake and herbage u t i l i za t ion information are required. ( i i ) P re - t r imming may not detect the decay of vegetation as the Difference Technique is able to do (Radcliffe, 1971). ( i i i ) There may be detrimental effects due to the tr imming cut. For example , the t r imming cut may affect species compos i t ion (Radcl i f fe , 1971) . App l i ca t ions of the Technique: Fuel leman and Bur l i son (1939 and 1940) used both the T r i m Technique and the Difference Technique to measure pasture yields and consumption under grazing conditions and to compare the yields and composi t ion of some I l l inois pasture plants. N o attempt was made to select the method that gave a more accurate representation of the actual y ie ld from a given pasture. Experience over several seasons of sampling seemed to indicate that the method used is dependent on the type of forage and the season. Unless a large number of samples was obtained on each sampling date, a number of negative yie ld figures may occur, especially in a system of random sampling. The T r i m method eliminated the presence of negative y ie ld figures regardless of the condit ions of the sward, the number of samples or the use of random samples. Thus the T r i m Technique did not detect the decay of vegetation as does the Difference Technique (Radcliffe , 1971). The yields of reed canarygrass , K e n t u c k y bluegrass (Poa pratensis L . ) and bromegrass, calculated from the trim and difference techniques, were found to vary depending on season and previous treatment of the pasture (Fuelleman and B u r l i s o n , 1940). There were no consistent differences between the 38 two methods. However , for orchardgrass the y i e l d calculated by the difference technique was greater than the y ie ld as calculated by the trim technique over three seasons of measurement. This may have been an effect of the t r imming cut reducing herbage y ie ld which would be more cr i t ica l for orchardgrass than for the other three grasses examined. There was no mention of the trimming height used in the trials. The Joint Commit tee (1943) made a prel iminary report on pasture investigation techniques for humid and irrigated sections of the U . S . . The report suggested that the tr im technique be used to measure the actual y ie ld or the seasonal growth of herbage. It was recommended that cages should not be placed on undergrazed areas of the pasture since the larger root reserves, or other condit ions, in the undergrazed grass may affect growth for a considerable period. The areas should be chosen carefully to secure representat ive vegetat ion and so i l type and to avo id recent droppings. It was suggested that at least 3 cages be used per site. N o recommendation on tr imming height was made. Nevens (1945) used the trim technique in a comparison of sampling procedures for making pasture determinations in I l l ino i s . The technique was characterized by smaller standard errors and lower coefficients of variabi l i ty than the difference techniques investigated. It was found to be advantageous in showing period yields and in comput ing yields where grazing is delayed because of rotational grazing or some other reasons. These advantages, together with its s impl ic i ty , make it highly suitable as the one plan of sampl ing where only y i e l d determinations of grass pastures are required. Radcl i f fe (1971) examined cutting techniques for pasture yields on different h i l l country sites in New Zealand. On North Island sites the trim technique consistently gave higher yields than the difference technique. On South Island sites, results showed no consistent overall effect, although large differences in yields were measured at some cuts. The number of samples required by the trim technique to give a standard error of 10% of the mean yield was 30 on North Island sites (sample size 0.3m x 0.5m). The difference technique was much more variable and required 5 and 20 times more samples on North and South Island sites respectively, to give a standard error of 10% of the mean yield. Variable ground cover within small areas made pairing of samples for the difference tephnique difficult resulting in higher variance. Variable growth for unlrimmed (compared to trimmed) grass also added to the variability. Radcliffe (1974) described the use of the technique in a series of experiments on grass-clover pastures to measure the seasonal growth pattern at various sites throughout New Zealand (Radcliffe, 1974 andl976; Baars, 1976a, 1976b; Baars et al., 1975). In these experiments the height of defoliation (2.5cm above ground) was chosen to simulate continuous sheep grazing under high stocking densities. The interval between defoliations (usually 2 to 3 weeks) was chosen as a reasonable, practical compromise between shorter staggered intervals which would give more precise estimates of the pasture growth rale, and longer intervals which in general would increase the recorded growth rate. The growth rates measured were related to the measurement technique used. There is not yet sufficient information to reliably extrapolate growth rates determined by one method to other defoliation methods (Radcliffe, 1974). Hayman and Moss (1979) examined the effect of winter management on pasture regrowth. They compared three systems of rotational grazing with strip grazing over two winters and measured pasture production in 4 0 the spring with the trim technique. The aim was to measure the effects of winter grazing management on spring pasture production. Radcl i f fe (1982) used the trim technique to examine the effects of aspect and topography on pasture product ion on h i l l country in N e w Zea land . Pasture measurements were obtained for a northerly and a southerly aspect containing several soil strata which were defined by their so i l profi le characteristics and topography. Herbage was harvested each month wi th handshears. A cage protected the regrowth and after each harvest was moved to a new position within 10 m of a previous sampling si te. The Pre -Tr im Technique for S low G r o w i n g Herbage: This technique was described by L y n c h (1947) and B r o w n (1954). L y n c h (1947) considered that this was the most satisfactory method adopted in N e w Zealand at that time. It is a technique that can be used in conjunction with a wide range of an imal trials. The method is used on pastures where the herbage growth, either at the beginning of the grazing season or at the end of a short grazing period, is lower than the cutting height which can be cut with a mower. The technique measures grassland production by using movable exclosures or cages which can vary in size according to the size of the exper imenta l f i e l d . De ta i l ed descr ip t ions of this method have been provided by L y n c h (1947), L y n c h and Mount ie r (1954) and E l l io t t and L y n c h (1958). Effects of the Tr imming Cut: Brown (1954) discussed the trimming cut and stated that a single tr imming does not change the botanical composition of the sward. However , t r imming is a drastic interference and results in an in i t ia l check on the growth rale. If it is assumed that in an annual plant the rate of increase of growth is proportional to the quantity of growth present, then the rate of growth in the trimmed, caged area at first lags behind that in the grazed area outside, but later, not being defoliated by graz ing , the caged herbage becomes taller than the grass outside and growth increases at a greater rate than in the grazed pasture (Brown, 1954). Frame (1984) considered the case of herbage inside the cage starting from a non-trimmed basis. Here the growth w i l l be greater than outside since it is not being defoliated l ike the outside; the overestimate w i l l be greater the longer the cage is down. In other words, sampling every 2 to 3 weeks is better than every 4 to 6 weeks. Conversely, to offset or correct for this undefoliated growth inside the cage you can pre-trim. The true estimate of the grazed sward growth w i l l l ie somewhere between the two but cannot be gauged accurately by caged methods, so the values obtained would not represent absolute values. Bosch (1956) averaged the non-t r immed and pre-tr immed results which is fair enough but it does seem somewhat arbitrary (Frame, 1984). Pre - t r imming sample sites, or t r imming post-grazed herbage evenly to sampling height w i l l markedly reduce the coefficient of var iabi l i ty of y ie ld measurements and herbage intake ('t Mannel jc , 1978; Frame, 1981). Nevens (1945) showed that the t r imming cut technique was characterized by smaller standard errors and lower coefficients of var iabi l i ty ( C V ) than the difference technique. Even though the t r imming cut does reduce the C V , this is not necessarily a jus t i f ica t ion for incorporat ing procedures w h i c h introduce an ar t i f ic ia l factor that may influence product ivi ty ('t Mannelje, 1978; Frame, 1981). Pre-tr imming is not recommended except where the practice is part of normal management ('t Mannetje, 1978). In the case of t ropical pastures containing t ra i l ing legumes, t r imming may result in a reduced legume content. L y n c h (1947) discussed his concerns with the t r imming cut. The pasture should be under control at all times. A lengthy growth when t r immed produces a form of recovery growth quite different from a pasture under normal grazing. It is not advisable to mow and weigh until at least a week after t r imming , since this regrowth grows at a rate different from normal growth and often has a different composit ion. Such growth tends to 'harden o f f in 3 to 4 days and become normal in al l respec ts . Frame (1981) cr i t ic ized the use of the t r imming cut on the grounds that herbage accumulation may in i t i a l ly be slower than on the untrimmed area outside the cage, but later this is reversed because of grazing on the untrimmed sward outside the cage. To work satisfactorily, the unprotected pasture should be defoliated to sampling height and this is very difficult to ach ieve . fc) Movable Cages - N o Trimming Cut: This technique was described by B o y d (1949) and Brown (1954). It involves the use of cages but without a tr imming cut. The technique was developed at the Rolhamsted Experimental Station, U . K Procedure ( B o v d . 1949) (i) the position of the cages is chosen at random. (ii) four cages per plot of 0.040 to 0.10 hectares are used. ( i i i ) herbage from under the same cage is cut from two parallel strips, A and B . The strips are 2.14m X 0.38m each. (iv) strip A is cut when the cage is put down and strip B when the cage is removed approximately 3 weeks later. (v) cages arc moved to a new location. Ahlgren et al. (1951) used a variation of the technique at the University of Wisconsin to compare the productivity of permanent and short term pastures. Movable wire cages were distributed at random over the fields. The forage beneath the cages was harvested in a manner to approximate the grazing at the end of each grazing period. Immediately following each harvest the cages were moved to new areas selected at r a n d o m . C a l c u l a t i o n s : The growth on an untrimmed strip is calculated from B - A . Average daily growth rates can be calculated by dividing the value, B - A , by the number of days the cage was at one position. If the area is sampled continuously over the growing season the seasonal distribution of growth rates can be calculated. Total seasonal production is obtained by summing the yield from each cage period for the growing season. Appl icat ion of the Technique: Thompson et al. (1976) used the version of Ahlgren et al. (1951) to measure pasture production from grazed paddocks in a study on the effects of stocking with rotational grazing on the productivity of dryland alfalfa in New South Wales, Australia. Percival and McClintock (1982) also used the method of Ahlgren et al. (1951) to measure pasture yields in a study examining the role of grazing management in manipulating the balance of ryegrass and paspalum in New Zealand pastures. 4 4 (cO Difference Techniques These techniques do not use a tr imming cut and are commonly used on the natural sheep pastures of Aus t ra l ia because of the severe effect t r imming has on herbage recovery (Brown, 1954). These techniques have been used at the I l l ino is Research Station and have been described by Fuel lcman and Bur l i son (1939 and 1940) and by Nevens (1941, 1943 and 1945 ) . General Procedure: (i) Pairs of quadrats are laid down on the pasture. ( i i ) One set of quadrats are on open pasture and are subject to grazing whi le the other set are covered by movable cages that protect them from grazing. ( i i i ) Both the open and closed quadrats are sampled at the same time. The interval between sampling can be at regular intervals of one, two or more weeks or sampling can be according to some stage of growth. Nevens (1945) sampled regularly at one month intervals. (iv) After sampling, the caged quadrats are moved to new locations. Nevens (1945) moved the cages a short distance to cover representative and not previously sampled herbage. Aus t ra l ian Di f fe rence Techn ique : A variation of the technique that was used in Aus t ra l ia (Lynch and Moun t i e r , l 954 ) . It is s imilar in some respects, but handshears or cl ippers are used to cut growth to ground level . After grazing finishes, the fol lowing cuts are made: (i) Herbage within the cages. Cages are then placed at new sites. (ii) 'Open Cuts' in the field on areas which have been grazed. The positions for these open cuts are found at random and the frames are placed on uncut areas adjacent to these open cuts. Calcula t ions: The amount of herbage consumed and lost by treading is measured by the difference between the open and caged quadrats sampled on the same day. The amount of growth is measured by subtracting the weight of the herbage (on a given date) sampled on an open quadrat from the weight of herbage on the adjacent caged quadrat that was sampled one, two or more weeks later. With measurements made over the entire growing season in this manner, both the seasonal distribution of growth and the total seasonal production can be obtained. With the Australian Difference Technique, herbage production is estimated from the amount cut from within the cage less that from open quadrats made when the cages were placed. These open cuts estimate the amount of herbage on the caged areas when the cages are placed in their new positions at the start of the measurement period: Herbage consumption is measured by the amount cut within the cage (amount available) less that of the open cut made on the same day (amount left after grazing). Applications of the technique: Robinson et al. (1937) compared grazing and clipping for determining the response of permanent pastures to fertilization. They used both the permanent cage method and the difference method. Fuelleman and Burlison (1939 and 1940) used the difference and trim methods to examine pasture yields and consumption under grazing conditions. In a comparison of the production obtained from 46 close-folding and rotational grazing of dairy cows, Holmes et al. (1950) used pre- and post-grazing strips without cages to measure herbage production. Under rotational grazing, herbage yield was measured by taking samples from open quadrats at the beginning of the grazing and from caged areas at the end of grazing. The caged areas measured the growth rate of ungrazed herbage during the grazing period. Campbell (1966) used the difference method to measure herbage production and availability under two different grazing systems. On rotationally grazed paddocks, pre- and post-grazing strips of herbage were taken and net pasture dry matter production corrected for during grazing by assuming that the growth rate during grazing is equal to the growth rate during the rest period. Under a grazing system of one day paddock and one night paddock, herbage production was measured by the difference method using movable cages. Waddington and Cooke (1971) used the technique to examine the influence of sample size and number on the precision of herbage production and consumption in two grazing experiments. In estimating herbage intake by grazing sheep, Walters and Evans (1979) compared the difference technique with several indirect animal techniques. It was found that the difference technique provided intake estimates in good general agreements with intake estimates obtained by animal methods, but discrepancies arose when used on a semi-prostrate variety, probably due to incomplete recovery of sampled herbage from post grazed quadrats. It was concluded that the difference technique might be usefully adopted for preliminary assessment of the intake characteristics of herbage varieties in routine evaluation programs, provided care is taken in collecting post-grazed samples and that a reduction in the land requirement can be achieved without undue loss of accuracy or precision. Lastly, Sharrow (1983) used the method to estimate forage standing crop and forage disappearance on rotationally grazed and continuously grazed pastures in Oregon. Forage dry matter intake by livestock was calculated from the difference between caged and uncaged plots on each sampling date using the equation of Linchan et al. (1952). 2.2.2.3 I N D I R E C T M E T H O D S Over the years several methods have been developed to estimate herbage mass while minimizing the cutting of herbage. Some of the reasons for these developments have been enumerated by Frame (1981). (i) T o reduce labour, equipment, time or resources needed and thus the cost of the measurement. (ii) T o make measurements on large fields or plots (particularly under grazing management) or on remote sites where it would be very difficult to sample swards by cutting. (iii) T o use in small scale grazing trials where sampling by cutting could affect a relatively large proportion of the treatment area. (iv) T o rank differences in trials with large comparative differences. (v) T o provide a guide to the estimation of herbage mass in animal production systems where an absolute measure is not necessary. 4 8 A general disadvantage of indirect techniques is that they assume a consistent re la t ionship between the indirect value and the actual dry matter production of the dry matter sampled (Boswe l l , 1980). M o s t indirect techniques use the double sampling concept ( W i l m et al . , 1944). The indirect measure and a cutting measure are used on a small number of samples and the re la t ionship between the two methods is derived and used to predict direct values from a large number of indirect m e a s u r e m e n t s . D i rec t estimates of herbage product ion are obtained by cutting herbage w h i l e i nd i r ec t est imates are based on severa l appra isa l techniques. A n y advantage to be gained from this approach depends on the relative cost and precision of double sampling when compared to direct sampl ing. M a n y indirect techniques termed non-destructive, do involve cutting of some herbage when the double sampling method is used. Several indirect techniques can be described, but this discussion w i l l be l imi ted to Weighted Di sc Methods. Other techniques not discussed in this review include: capacitance meters, visual assessments, height/density m e a s u r e m e n t s , r emote s e n s i n g , gas e x c h a n g e , be ta -a t t enua t ion , photography, compressibi l i ty and measurement of leaf area. W e i g h t e d D i s c M e t h o d s : Sward height and density are the two main characteristics that influence herbage mass and its visual appraisal. Height and density have both been used, separately or in combination, in a large number of techniques. B r o w n (1954) reviewed earlier methods based on the re la t ionship between sward height and density and herbage mass. M o r e recent reviews have been conducted by 't Mannetje (1978), Frame (1981) and Mei js et a l . (1982). Height is normally defined as maximum or mean and is measured by a ruler. Density is defined as percent ground cover and is estimated by point quadrat or visual appraisal (Bakhuis, 1960). Simple weighted disc instruments have been developed to estimate herbage mass. The disc normally consists of a light aluminum plate fitted over a shaft held in a vertical position with its base at the ground. Other materials from which the disc can be constructed are: wood, plastic or polystyrene. The disc may be allowed to settle to a constant position on the sward or the stems may be pushed down onto the ground as the disc lies on the sward. The instrument readings are influenced by a combination of sward height and density. Readings are calibrated against herbage mass measured by cutting. There arc inherent difficulties in estimating pasture height and ground cover and combining them into a single expression. Sullivan et al. (1956) first overcame these difficulties by using a board technique in which a plywood board was placed on top of the herbage to effectively integrate height and density into a single expression. The final resting height of the board gave a measure of the bulk density of the pasture (Sullivan et al. , 1956 and Shrivastava et al., 1969). Jagtenburg (1970) proposed a simple technique for estimating herbage mass by placing a rigid weighted sheet on the grass surface and measuring the average distance to the ground after some settling period. Design and operation of weighted disc meters: Phillips and Clark (1971) developed a weighted disc apparatus in New Zealand, called the Phillips Disc Grassmeter, which is manufactured commercially. The meter consists of a 35.6 cm diameter metal disc that can be raised or lowered 5 0 wi th in a frame formed by three inverted L-shaped legs. The legs are placed at a site where a reading is required and the disc lowered onto the pasture. After 10 seconds of 'settling time' a small brake is applied to the vertical shaft carrying the disc. A tape measure attached to the disc and to the other end of the vertical shaft records the height in centimeters above the soil level to which the disc is raised by the herbage. A s imple ins t rument was deve loped at the Hannah Research Institute, U . K . in 1969 and tested regularly for a number of measuring purposes. Castle (1976) described and evaluated the instrument. It has a shaft and two l inked discs (Figure 2). The aluminum shaft is marked in subdivisions of 0.50 cm. A pin, riveted through the shaft, holds the bottom disc in place when the instrument is moved. The two discs are made from a luminum sheet (20 gauge) and, inc luding the three l i nk ing rods, weighs 200 grams. The top disc is painted matte red to prevent reflection. When in use, the shaft is held in a vertical position with its base on the ground, and the discs are placed over the shaft with the smal l disc on top. The large lower disc settles to a constant posi t ion on the herbage and this height is read from the position of the small disc to an accuracy of 0.5 cm (Figure 3). M c C a r t h y and Kearney (1977), work ing in Ireland, evaluated the 'Grassmeter' produced commerc ia l ly in the U . K . , which consisted of an aluminum base plate sl iding on a calibrated bar. In their assessment of four techniques for est imating y ie ld on dry land pastures, M i c h a l k and Herbert (1977), used a p l y w o o d board based on the idea of Sul l ivan et al . (1956). A 3.1 mm X l m ^ p l y w o o d boa rd weighing about 2 kg was dropped from a height of about 1.2 m onto the pasture at random locations. The vert ical height of the board above the ground was measured through a hole at the center of the board. Ea r l e and M c G o w a n (1979) evaluated an automated r i s ing plate meter which they cal led the El l inbank Pasture Meter ( E . P . M . ) that was an automated version of the Massey Grass Meter (Holmes, 1974). When the handle is held by the operator and the other end of the shaft is rested on the ground, pasture prevents the plate from fal l ing to ground level and in effect raises the level of the plate. W i t h the E . P . M . the height is not observed at each locat ion but a cumulat ive measure of the height is recorded on a counter. When the meter is lifted from the pasture, the plate drops down the shaft to rest on a bottom washer and push a wire to activate a second counter and record that a reading has been made. For a given pasture, the height at which the plate is held is dependent on the height and density of the pasture and so w i l l indicate pasture y ie ld . The quantitative relat ionship between plate height and pasture y ie ld w i l l be affected by weight per unit area of the plate. Earle and M c G o w a n (1979) used 4.0 k g / m ^ in 1975 and 1976 but increased the density of the disc to 5.0 k g / m ^ in 1977 to al low for an additional weight of a second counter and to provide greater overall strength to the meter. A d v a n t a g e s : (i) Weighted disc meters can be used on commerc ia l farms or in advisory work. ( i i ) Speed and s impl ic i ty of sward height observations can be an advantage in a l l owing a large number of measurements to be made on non-uniform grazed swards. ( i i i ) Weighted disc instruments are easily dupl icated and require little s k i l l for operation. A n operator can be trained in a short period. ( iv) Weighted disc instruments are inexpensive, simple to construct and need l i t t le repair which are advantages when compared to electronic pasture meters. Di sc meters do not suffer calibration shifts due to battery, temperature or humidity fluctuations. Wetness of pasture does not affect disc meters. They are easier to transport and do not suffer calibration shifts due to transport as is possible with capacitance meters. Disadvantages : (i) D i s c meters have sources of bias which include: sward structure, lodged or trampled herbage, botanical compos i t ion , season and grazing m a n a g e m e n t . ( i i ) D i s c meters often need recalibration because of differences in regressions for different seasons and different sward types. ( i i i ) They are difficult to use on uneven or muddy pastures. ( iv) They have not been evaluated under a variety of grazing s i tua t ions . (v) The i r su i tab i l i ty on heterogeneous pastures or rangelands is l imi ted (Tucker, 1980). Some applications of weighted disc meters: (i) Used to quickly determine cutting date in plot trials. ( i i) Used when yields of uninterrupted growth are required. ( i i i ) Can be used to determine the approximate date when herbage begins to grow in the spring (Jagtenberg, 1970). (iv) Used in grazing experiments when animals should begin grazing at a certain dry matter y ie ld . (v) Used to determine the amount of material on offer to l ivestock (Castle and Watson, 1973). (vi) Usefu l in farm management and extension work . In N e w Zealand, Phi l l ips and Clark (1971) reported that the disc meter is a real aid to pasture management. FIGURE 2 ; DETAIL OF DISC METER 56 I T T E X P E R I M E N T A L M A T E R I A L S A N D M E T H O D S 3.1 A G A S S I Z C U T T I N G T R I A L : The Experimental Site was located at the Agriculture Canada Research Station, Agassiz, British Columbia on a Degraded Enlric Brunisol / Gleycd Degraded, Melanic Brunisol soil of medium texture. The area was seeded by hand with two orchardgrass varieties: Prairial (medium) and Sumas (late) at a rate of 33 kg/ha and two diploid perennial ryegrass varieties, Norlea (late) and Barlano (early) at a rate of 39 kg/ha. The seedbed was fertilized to soil test requirements and the area sprayed with herbicide to control weeds. Treatments : Two orchardgrass varieties ( Dactylis glomerala L . ) Prairial and Sumas and two diploid perennial ryegrass varieties ( Lolium perenne L.) Norlea and Barlano were examined in the experiment. Each variety was subjected to two cutting regimes. These cutting regimes were: Low Infrequent ( LI): Cut to a height of between 2.5 cm to 4 cm and growth allowed to occur for 4 to 6 weeks before the next harvest. High Frequent (HF): Cut to a height of between 7.5 cm to 10 cm and growth allowed to occur for 3 to 4 weeks before the next harvest. Exper imenta l Des ign : The experiment was conducted using a randomized block design. The four grass varieties were arranged in main plots (2.7 X 6.1m). Cutting treatments were arranged as subplots (1.35 X 57 6.1m) . A l l possible treatment combinations were replicated four times. A 0.61m aisle was located between each block and the entire area was surrounded by a 3m border area. A l l aisles and border areas were kept mown at all times (Figure 4). Under the H F cutting regime, eight cuts were obtained for the two orchardgrass varieties in 1984 and 1985 whereas only seven cuts were obtained for the two ryegrass varieties in both of these years. In 1984 both ryegrass varieties did not have sufficient regrowth after the seventh harvest to make an eighth cut possible. In 1985 both ryegrasses experienced very depressed growth during mid-summer and on one occasion for each variety there was not enough growth to harvest. Table 1 outlines harvest dates for 1984 and 1985. Under the L I cutting regime, six cuts were obtained for Prairial, Sumas and Barlano in 1984 and 1986. Only five cuts were obtained for these three varieties in 1985 due to the low rainfall over the summer and the lack of regrowth needed for a sixth and final cut. Six cuts were obtained for Norlea in 1986 but only five cuts were obtained in 1984 and four cuts in 1985. Norlea proved to be susceptible to dry summer weather and no growth was available for harvest in August of 1984 and 1985. Table 2 outlines harvest dates for 1984, 1985 and 1986. Plot harvesting and fertil ization: A l l plots were harvested with a 'Swift Current' plot harvester equipped with an adjustable bar under the blades that allowed cutting height to be varied. Cutting width of the mower was 0.61m. Yields of fresh herbage were determined at every cut in the season by cutting two swaths each 0.61 by 5.5m down to the appropriate cutting height on each plot. Each swath was cut twice; once in TABLE 1: HARVEST DATES FOR HIGH - FREQUENT CUTTING TREATMENT, AGASSIZ - 1984 AND 1985 1984 ALL FOUR MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 CULTIVARS 1985 ALL FOUR MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 CULTIVARS TABLE 2: HARVEST DATES FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ - 1984. 19( YEAR AND HARVEST CULTIVAR 1 2 3 1984 PRAIRIAL MAY 8 JUNE 5 JULY 3 AUGUST 2 SEPTEMBER 7 OCTOBER 12 SUMAS MAY 8 JUNE 5 JULY 3 AUGUST 2 SEPTEMBER 7 OCTOBER 12 BARLANO MAY 8 JUNE 5 JULY 3 AUGUST 22 SEPTEMBER 21 OCTOBER 12 NORLEA MAY 8 JUNE 5 JULY 3 AUGUST 22 SEPTEMBER 21 1985 ALL FOUR MAY 6 JUNE 4 JULY 3 AUGUST 5 SEPTEMBER 20 CULTIVARS 1986 ALL FOUR MAY 7 JUNE 7 JULY 8 AUGUST 8 SEPTEMBER 19 OCTOBER 25 CULTIVARS F I G U R E 4: F I E L D P L A N O F A G A S S I Z C U T T I N G T R I A L 0, KI HI (L KI I? IL KI KI g KI IL IF KI IL D KI ] IL i IL i KI us KI I IF KI IL % KI 1 IL E KI 9 L I KI OK KI IL % U \ h IL KI Key: T w o pGrennial r yeg ras s c u l t i v a r s : B=Barlano, N=Norlea. T w o o rchardgras s c u l t i v a r s : P =P r a i r i a l , S=Sumas. H=High-Frequent cu t t i n g and L=l_ow-lnfrequent c u t t i n g 60 either direction. Herbage from the plots was transferred from the harvester bin onto a tarpaulin and then into plastic bags. Herbage was weighed in these bags to the nearest 50 g on electronic scales in the building adjacent to the experimental site. Two random sub-samples of herbage weighing approximately 1 kg were immediately taken for each plot and sealed in plastic bags. Samples were then transported to the Animal Science Laboratory at U . B . C . One sub-sample was used for dry matter ( D . M . ) determination by drying a 300 to 400g sample in an aluminum pan in a forced draught oven at 1 0 0 ° C for 24 hours. The second sub-sample was dried at 6 0 ° C for approximately 48 hours and subsequently used for proximate analysis. A n y unsampled herbage remaining on the plots was cut to the appropriate height and discarded. A l l plots were fertilized in Apri l of each year according to soil test recommendations made in the fall of the previous year conducted by Agriculture Canada. Subsequent fertilization was made on a monthly basis thereafter, with all plots receiving fertilizer at the same time. A total of 300kg/ha of nitrogen was applied each year. A l l fertilizer was broadcast in granular form by hand as ammonium nitrate, ammonium sulphate or an 18-18-18 ( N - P - K ) mixture with urea or monammonium phosphate as the nitrogen source. Plots were limed with agricultural lime in May 1985 at a rate of 2 tonnes per hectare. For the duration of the trial no herbicides, insecticides or irrigation were applied to the plots. Botanical analysis and percent ground cover were determined by the point quadrat method with 25 measurements per quadrat and four quadrats placed randomly on each subplot. Measurements were taken in the spring and fall of each year. 3.2 A B B O T S F O R D C U T T I N G T R I A L : The experimental site was located at the Abbotsford Test Station, Abbotsford, B . C . of the B . C . Ministry of Agriculture and Fisheries on moderately coarse to coarse textured eolian deposits. Plots were laid out on an existing orchardgrass / perennial ryegrass sward of undetermined age which had been utilized for several years as a hay field. Poultry manure had been applied frequently over the years and consequently the sward had a high infestation of chickweed as well as dandelions. The trial under discussion was conducted from May to October 1985. T r e a t m e n t s : Treatments consisted of two cutting heights and two cutting intervals. These cutting regimes were: L o w Infrequent (LI): Cut to a height of between 2.5 cm to 4 cm and growth allowed for 4 to 6 weeks before the next harvest. L o w Frequent (LF) : Cut to a height of between 2.5 cm to 4 cm and growth allowed to occur for 3 to 4 weeks before the next harvest. High Infrequent (HI): Cut to a height of between 7.5 cm to 10 cm and growth allowed to occur for 4 to 6 weeks before the next harvest. High Frequent (HF): Cut to a height of 7.5 cm to 10 cm and growth allowed to occur for 3 to 4 weeks before the next harvest. E x p e r i m e n t a l D e s i g n : The experiment was conducted using a randomized block design. The four cutting treatments were arranged , in 62 F I G U R E 5 : F I E L D P L A N O F A B B O T S F O R D T R I A L S ( a ) C u t t i n g T r i a l ,(b) F e r t i l i z e r T r i a l ILO ILIF DD DDD D¥ D UF MD MIF [LD D DW DD DDD MIF LU n MB ILIF OV DD D DDD MD ILtF •IF ILD W DDD D DD K e y - L l = L o w I n f r e q u e n t c u t t i n g K e y : F e r t i l i z e r r e g i m e s I, II, III & IV. HI = H i gh I n f r e q u e n t c u t t i n g LF = L o w F r e q u e n t c u t t i n g HF = H i g h F r e q u e n t c u t t i n g 63 main plots (2.44 X 6.1 m). A 0.61 m aisle was located between each block and the entire area was surrounded by a 3m border area. A l l aisles and border areas were kept mown at all times. A l l possible treatment combinations were replicated four times (Figure 5). Plot Harvesting and Fertilization: A l l plots were harvested with a 'Swift Current' plot harvester in an identical manner to the Agassiz cutting trial. Samples and subsamples were also treated similarly. Plots were fertilized in early May according to soil test recommendations made from soil analyses conducted by the B . C . Ministry of Agriculture and Fisheries. Subsequent fertilization was made at monthly intervals with a total annual application of 300 kg Nitrogen/ha. A l l fertilizer was broadcast by hand in granular form and the types of fertilizer and nitrogen sources were the same as for Agassiz. For the duration of the trial no lime, herbicides, insecticides or irrigation were applied to the plots. 3.3 A B B O T S F O R D F E R T I L I Z E R T R I A L : The experiment was located adjacent to the plots described above in the Abbotsford cutting trial. Soil and pasture type were very similar. T r e a t m e n t s : The treatments consisted of varying fertilizer application. There were four treatments which are outlined in Table 3. E x p e r i m e n t a l D e s i g n : The experiment was conducted using a randomized block design. The four fertilization treatments were arranged as main plots ( 2.7m X 6.1m ). A 0.61 m aisle was located between each block and the entire area was surrounded by a 3m border. Aisles and 64 borders were kept cut at all times. A l l possible treatments were replicated four times (Figure 5). T A B L E 3: Abbotsford fertilizer trial treatments. Date of Application Treatment (applied N in kg/ha) I II I l l I V May 22 150 100 100 June 16 100 50 100 July 12 50 50 50 August 22 - 100 50 Total Appl ied (kg N/ha) 0 300 300 3 0 0 Plots were initially intended to be harvested at monthly intervals, but due to a very dry summer, growth was greatly reduced in August and September resulting in a longer period between harvests. Plots were harvested according to the following schedule: May 22, June 17, July 12, August 22 and October 5. A l l plots were harvested with the Swift Current plot harvester in the same manner as the Agassiz cutting trial. Samples and subsamples were treated in a similar manner as those of the Agassiz Trial . Soil test results were obtained from the B . C . Ministry of Agriculture and Fisheries and applied fertilizer was in the same form as the Agassiz trial . 65 3,4 OYSTER RIVER FERTILIZER TRIAL: The Experimental Site was located at the U . B . C . Farm #2, Oyster River, B . C . . Plots were laid out on an existing sward consisting of orchardgrass, perennial ryegrass, white clover and red clover. The field was sown in 1982 and had been utilized for silage, hay, and grazing. The trial under discussion began in A p r i l 1985 and continued until October 1986. However, only data for the first year are presented. T r e a t . i n en ts: The experiment consisted of varying fertilizer applications. There were four fertilization regime treatments which are outlined in Table 4. Table 4: Oyster River Fertilization Trial Treatments Date of Application Treatment (applied N in kg/ha) I I I I I I I V Apri l 11 1 0 0 100 1 0 0 10 0 May 9 0 1 25 100 50 June 10 0 75 50 50 July 8 0 0 50 0 August 12 0 0 0 100 Total Applied (kg N/ha) 100 3 00 300 300 E x p e r i m e n t a l D e s i g n : The experiment was conducted using a randomized block design. The four fertilizer treatments were arranged as main plots (3.05 x 6.10m). A 0.61 rn aisle was located between each block 66 FIGURE 6: FIELD PLAN OF OYSTER RIVER FERTILIZER TRIAL 1 III II IV II IV 1 III IV III II 1 III 1 IV II K e y : F e r t i l i z a t i o n reg imes I, II, III and IV. 67 and the entire area was surrounded by a 3.6m border (Figure 6). A l l aisles and border areas were mown at monthly intervals when the samples were collected. A l l plots were sampled with a Swift Current plot harvester and samples and subsamples were treated in a similar manner to those from the Agassiz cutting trial. Harvest dales were: May 9, June 10, July 8, August 12, September 11 and October 29. A l l plots received the same amount of potash and phosphate fertilizers. Types of fertilizer and nitrogen sources were the same as for Agassiz. Soil test results were obtained from the U . B . C . Farm #2 soil sampling information. For the duration of the trial, no lime, herbicides or pesticides were applied to the plots. Irrigation was applied as part of farm management in July, but no records of how much water was applied are available. Botanical analysis was conducted in the same manner as measurements at Agassiz. 3.5 F O R A G E Q U A L I T Y : Samples were prepared for analysis by using one of the sub-samples from each plot. A 300 to 400g sample was dried at 6 0 ° C for about 48 hours and subsequently ground to pass through a 1 mm screen. Ground samples from each treatment replicate were then composited and stored in plastic bags. Samples were then analyzed for the following: Nitrogen (N), acid detergent fibre ( A D F ) , ash, calcium, phosphorus, potassium, sodium, magnesium, manganese and copper. N i t rogen: Determination of percent N was obtained by using the digestion method of Parkinson and Allen (1975) in which organic matter N was converted to ammonium sulphate. After digestion, the solution was measured for ammoniacal nitrogen content using the Technicon Auto 68 Analyzer II (1974) using the technique developed by Wal l and Gehrke (1975). A c i d Detergent Fibre: A D F content of herbage samples was determined by the method of Waldern (1971). A s h : Herbage samples were analyzed for ash content by combusting samples for 24 hours at 6 0 0 ° C . M i n e r a l s : A l l minerals, except phosphorus, were analyzed by atomic absorption using the acid digest from the N determination (Heckman, 1967; Christian and Feldman, 1970; Perkin Elmer, 1976). Phosphorus was determined by analyzing the acid digest on the Technicon Autoanalyzer II (1974). A l l glassware used for sample preparation was acid washed to remove mineral contamination. 3.6 D I S C M E T E R I N V E S T I G A T I O N S : Three weighted disc meters were constructed according to the design of Castle (1976) (Figures 1 and 2). When measuring herbage, the shaft was held in a vertical position with its base on the ground and the discs were placed over the shaft with the small disc on top. The large lower disc was allowed to settle to a constant position on the herbage. After allowing for a settling time of 10 seconds, the height of the lower disc above ground level was read from the top disc to an accuracy of 0.5cm. Although the position of the disc is determined by grass height, grass density and the proportion of stem in the herbage, for ease of expression all measurements will be referred to as 'height' The instrument was tested against known yields of D . M . under the following conditions: (i) A grazed dairy pasture at Agriculture Canada's Farm #2, Agassiz. (ii) A grazed dairy pasture in Langley, B . C . (iii) Pure swards of Prairial and Barlano at Agassiz. Usually 10 or 20 measurements were taken at each site on any particular sampling date. After each height measurement was recorded, the discs were left in position and a plywood template placed over the lower disc and secured into position. The discs were then removed and the herbage inside the template harvested to ground level with electric sheep shears. This material represented all the herbage upon which the disc had settled. Herbage samples were placed in paper bags and dried immediately at 1 0 0 ° C for 24 hours in a forced draught oven. Dry matter yield (kg DM/ha) was then correlated to disc 'height' for each site and sampling date. I V R E S U L T S A N D D T S C U S S T O N 4.1 CLIMATIC DATA As with any year in which climatic data is examined the records over the three years of the trial at Agassiz showed considerable variation between years and with the 96 year long term average. Figures 7 to 10 show the total monthly precipitation and average monthly minimum and maximum air temperatures. Monthly precipitation was well above average during January, May, September and November of 1984 with only June being lower than average (Figure 10). Annual precipitation during 1984 was 31% higher than the long term average. In contrast, 1985 had 12% lower rainfall than the long term average (67% of 1984 rainfall). January, February and March, 1985 had below average rainfall while April had twice average values. High rainfall in Apri l resulted in lower than normal soil temperatures delaying the onset of spring growth. The months of July and August were not only exceptionally dry (2.6mm and 25.8mm respectively) but July had higher than average air temperatures and August was at average values. Annual precipitation for 1986 was 11% greater than the long term with higher rainfall during February, May, June and November. August was the only growing month receiving below average rainfall (5.0mm). In each of three years of the trial the herbage was always moisture stressed during July and August; this was especially evident in 1985. Monthly averages can be somewhat misleading since they do not indicate extended periods without precipitation. July and August were both relatively dry months in 1984 and 1986, with most of the rain falling F I G U R E 7 : A V E R A G E M O N T H L Y T E M P E R A T U R E . A G A S S I Z 1964 30 L U - 1 0 1 ' • • • • 1 •—. 1 i i i I JAN FEB MAR APR MAY JUNE JULY AUG SEPT OCT NOV DEC F I G U R E 8 : A V E R A G E M O N T H L Y T E M P E R A T U R E . A G A S S I Z 1985 JAN FEB MAR APR MAY JUNE JULY AUG SEPT OCT NOV DEC 72 FIGURE 9: AVERAGE MONTHLY TEMPERATURE. AGASSIZ 1986 JAN FEB MAR APR MAY JUNE JULY AUG SEPT OCT NOV DEC FIGURE 10; MONTHLY PRECIPITATION. AGASSIZ 1984.1985 AND 1906 400 JAN FEU MAR APR MAY JUNE JULY AUG SEPT OCT NOV DEC 7 3 during the last week of August, 1984 and during the first two weeks of July, 1986. 4.2 A G A S S I Z C U T T I N G T R I A L 4.2.1 A n n u a l dry matter yields: Annual dry matter production averaged over three years for all four cultivars under the low-infrequent (LI) cutting treatment is shown in Table 5. There were significant (P<0.05) yield differences between cultivars with the following ranking: Prairial > Sumas > Barlano > Norlea. The same ranking was obtained for the annual production averaged over two years for all four cultivars under the high-frequent (HF) cutting treatment (Table 6). Again , yield differences between all four cultivars were significantly (P<0.05) different. Relative yields between cultivars for both cutting treatments were very similar. With the yield of Prairial assigned an index value of 100, the relative yield of: Sumas was 90 for LI cutting and 92 for H F cutting; Barlano was 80 for LI cutting and 82 for H F cutting and Norlea was 63 for both LI and H F cutting treatments. Although yields were slightly less under H F cutting than LI cutting, the ranking of cultivars and relative yields of cultivars within a cutting treatment were unaffected by cutting treatment. Examination of annual yields under L I cutting for 1984, 1985 and 1986 (Table 5) and H F cutting for 1984 and 1985 (Table 6) revealed the same yield ranking of cultivars in all cases. In general the orchardgrass cultivars (Prairial and Sumas) yielded more than the perennial ryegrass cultivars (Barlano and Norlea) with Prairial yielding more than Sumas and Barlano yielding more than Norlea. Annual yields were similar to those obtained by Fairey (1985a) at the same location during 1979 and 1980. Two orchardgrass cultivars (Saborto and Tendprbile) and two perennial ryegrass cultivars (Cropper and Barlatra) were examined under a 4 cut management system with an annual application of 300kg N/ha . Orchardgrass (average yie ld 1 1394kg D . M . / h a ) yielded more than perennial ryegrass (average yield 10148kg D.M. /ha) . Average yields in this study are slightly lower than those of Fairey (1985a) mainly due to the low yields obtained in 1985. The yield of Barlano perennial ryegrass during 1984 and 1986 was also similar to yields of eight perennial ryegrass cultivars evaluated at the same location by Fairey (1985b) where dry matter yields averaged 9180kg D . M . / h a from 1979 to 1981 under 4 cut (300kg N/ha) and 8 cut (450kg N/ha) management systems. The older cultivar, Norlea, yielded less dry matter than Barlano, a finding in agreement with local variety testing (Bates, 1986) Both Prairial and Sumas produced yields superior to those reported by Davis (1961) who evaluated 40 orchardgrass varieties at the same location during 1957 to 1959. Yields of the belter cultivars averaged 5600 to 6120 kg D . M . / h a . Such low comparative yields are probably attributable to the lower levels of nitrogen fertilizer used (97 kg/ha/year) compared to that used in this study (300 kg/ha/year). Local variety testing indicates similar relative yields between Prairial and Sumas. During 1983 and 1984 Sumas had relative yields of 94 at Agassiz and 92 at Sumas Prairie (approximately 30 km from the test site) when compared to Prairial (Bates, 1986). Average annual yields for both cutting treatments over all harvest years are shown in Table 7. Average annual yields under LI cutting ,for 1984 and 1986 were almost identical. Significant yield reductions occurred during 1985 under both cutting treatments for all four cultivars with yield reduced to about 58% of 1984 levels. This yield reduction can be attributed to both the cool, wet spring in 1985 which reduced spring growth when compared to 1984 and the extremely long, dry summer which greatly reduced summer yield when compared to 1984 and lowered fall regrowth and yield. Individual cultivar yields all showed significant (P<0.05) yield reductions in 1985. For the LI cutting, cultivar yields in 1984 and 1986 showed no significant (P>0.05) difference with the exception of Norlea which had increased yield in 1986 compared to 1984. Analysis of variance for average annual yields of both H F and LI cutting treatments revealed a slightly significant orchardgrass vs. perennial ryegrass x year interaction (P<0.05) which implies that annual yield difference between orchardgrass and perennial ryegrass changed over the seasons of measurement. For both H F and LI cutting treatments the relative yield of Sumas to Prairial remained the same over all seasons (Tables 5 and 6). However, the relative yield of Barlano increased in 1985, moving from an index of 76 in 1984 to 84 in 1985 for LI cutting and from an index of 78 in 1984 to 88 in 1985 for H F cutting. The relative yield for Norlea did not change between 1984 and 1985 for L l cutting and actually decreased somewhat for H F cutting. These results indicate that although yields were significantly reduced in 1985 the yield reduction of orchardgrass was comparatively greater than that of Barlano ryegrass. The relative increase in yield for Norlea in 1986 under LI cutting is somewhat misleading because the botanical composition of Norlea plots changed appreciably over the duration of the trial. Table 8 indicates the botanical composition of all four cultivars under both cutting treatments. TABLE 5 : ANNUAL AND THREE YEAR AVERAGE YIELDS FOR LOW-INFREQUENT CUTTING TREATMENT, AGASSIZ - 1984. 1985 AND 1986 CULTIVAR ANNUAL YIELD (kg DM/ha) THREE YEAR AVERAGE (with relative yield for each year) ANNUAL YIELD 1984 1985 1986 (kg D.M. / ha) PRAIRIAL 12,739 a 100 7,210 a 100 11,986 a 100 10,644 a 100 SUMAS 11,320 b 89 6,466 a 90 10,825 b 90 9,537 b 90 BARLANO 9,623 c 76 6,074 a 84 9,649 c 81 8,448 c 80 NORLEA 7,328 d 58 4,186 b 58 8,487 d 71 6,666 d 63 MEAN 10,252 5,984 10,236 8,824 S.E. OF MEAN 308 562 199 221 L.S.D. 697 1,272 450 500 Note: Means within the same column with different letters are significantly different (P=0.05). TABLE 6 ; ANNUAL AND TWO YEAR AVERAGE YIELDS FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ - 1984 AND 1985 CULTIVAR ANNUAL YIELD (kg DM/ha) TWO YEAR AVERAGE (with relative yield for each year) ANNUAL YIELD 1984 1985 (kg D.M. / ha) PRAIRIAL 11,924 a 100 6,857 a 100 9,390 a 100 SUMAS 10,947 a 92 6,306 a 92 8,625 b 92 BARLANO 9,310 b 78 6,061 a 88 7,686 c 82 NORLEA 7,737 c 65 4,170 b 61 5,953 d 63 MEAN 9,979 5,848 7,914 ,E. OF MEAN 535 446 334 L.S.D. 1,211 1,009 756 Note: Means within the same column with different letters are significantly different (P=0.05). —i TABLE 7 : ANNUAL YIELDS UNDER LOW-INFREQUENT AND HIGH-FREQUENT CUTTING TREATMENTS AGASSIZ. 1984-1986 HARVEST YEAR PRAIRIAL CULTIVAR SUMAS BARLANO NORLEA LOW-INFREQUENT CUTTING 1984 12,739 a 11,320 a 9,623 a 7,328 b 1985 7,210 b 6,466 b 6,074 b 4,186 c 1986 1 1,986 a 10,825 a 9,649 a 8,487 a LSD (P=0.05) = 843 SE of Mean = 390 HIGH-FREQUENT CUTTING 1984 1985 LSD (P=0.05) = 1061 11,924 a 10,947 a 9,310 a 7,737 a 6,857 b 6,306 b 6,061 b 4,170 b SE of Mean = 478 CUTTING TREATMENT LOW-INFREQUENT HIGH-FREQUENT 1984 10,252 a 9,979 a 1985 5,984 b 5,848 b 1986 10,237 a -MEAN 8,824 7,914 SE of Mean 393 483 L.S.D. (P=0.01) 1,100 1,476 cc FIGURE 11: ANNUAL YIELD AT AGASSIZ LOW - INFREQUENT CUTTING, 1984 TO 1986. FIGURE 12: ANNUAL YIELD AT AGASSIZ HIGH - FREQUENT CUTTING, 1984 AND 1985 YIELD (kg D.M./ha) (Thousands) PRAIRIAL SUMAS BARLANO NORLEA CULTIVAR i i H 1985 III 1984 I FIGURE 13: ANNUAL YIELDS AT AGASSIZ LOW-INFREQUENT AND HIGH-FREQUENT CUTTING 1984 AND 1985 YIELD (kg D.M./ha) (Thousands) P - Ll P - HF S - Ll S - HF B - Ll B - HF N - HF N - Ll CULTIVAR & CUTTING TREATMENT H 1985 H I 1984 LI-LOW-INFREQUENT & HF-HIGH FREQUENT P - PRAIRIAL, S • SUMAS B • BARLANO, N - NORLEA 82 A t the beginning of the trial all plots were pure stands of the seeded cultivars and as the trial progressed some weed infestation occurred since no herbicides were used. Only Norlea suffered from any significant weed encroachment and this was more severe under LI cutting (Table 8). This can be attributed in part to Norlea's growth reduction/dormancy during dry summer periods (which is exacerbated by low cutting) that offers little competition for weeds. It should also be noted that about 18% of the herbage in Norlea plots consisted of orchardgrass which contributed to the dry matter yield and which accounts for a significant portion of the relative yield increase for Norlea when compared to Prairial under LI cutting for 1986. Visual observations of the plots in 1985 showed no orchardgrass contamination of Nor lea plots and much less weeds suggesting , that the very dry summer period of that year could have weakened the Norlea stand to allow orchardgrass and weeds to establish. Table 8: Botanical Composition of Agassiz plots.* Cul t ivar Percent weed infestation L o w - I n f r e q u e n t H i g h - F r e q u e n t Nor lea Bar lano Sumas P r a i r i a l 24 (18% orchardgrass) 1 4 5 4 4 6 4 5 *Measured in May 1986. TABLE 9 : ANNUAL YIELD FOR BOTH CUTTING TREATMENTS AGASSIZ - 1984, 1985 and 1986. CULTIVAR CUTTING ANNUAL YIELD (kg D.M./ha) TREATMENT 1984 1985 1986 PRAIRIAL LOW-INFREQUENT 12,739 a 7,210 a 11,986 a PRAIRIAL HIGH-FREQUENT 11,924 ab 6,857 a -SUMAS LOW-INFREQUENT 11,320 be 6,466 a 10,825 b SUMAS HIGH-FREQUENT 10,947 c 6,306 a -BARLANO LOW-INFREQUENT 9,623 d 6,074 a 9,649 c BARLANO HIGH-FREQUENT 9,310 d 6,061 a -NORLEA LOW-INFEQUENT 7,328 e 4,186 b 8,487 d NORLEA HIGH-FREQUENT 7,737 e 4,170 b -MEAN 10,116 5,916 10,236 S.E. OF MEAN 413 491 199 L.S.D. 914 1,103 450 Note: Means within the same column with different letters are significantly different (P=0.05) 84 4.2.2 A n n u a l Dry Mat ter Yie ld - Cut t ing Treatment Effects: Cutting regime did not have a significant effect on annual yield. There were no significant (P>0.05) differences between annual yields under H F and LI cutting for each cultivar (Table 9). Annual yields were affected to the greatest extent by year to year climatic variables, followed by cultivar differences and to the least extent by cutting regime. Average annual yields over all four cultivars showed no significant yield advantage for either cutting regime (Table 10). However, there was a tendency for LI cutting to produce higher yields than H F cutting for Prairial, Sumas and Barlano. The reverse is true for Norlea where H F cutting produced slightly higher annual yields probably due to the fact that LI cutting had a deleterious effect on Norlea growth during the dry summer months. Table 10: Comparison of Agassiz cutting treatments. Cutting Treatment Y e a r 1984 1985 L o w - I n frequent 10,252 a 5,984 a H i g h - F r e q u e n t 9,979 a 5,848 3 M e a n 10,116 5,916 S.E. of Mean Not sig. Not sig. L.S.D. Not sig. Not sig. Note: Means within the same column with different letters are significantly different 4.2.2.1 Cutting Frequency: It is generally noted that less frequent defoliation results in higher yields (Woodman and Norman, 1932; Burger et al.,1958; Lambert,1962 85 and Bland, 1967). Anslow (1967) worked with perennial ryegrass and showed that herbage production was higher throughout the season (April to November) on a longer cutting interval of 6 weeks compared to 3 weeks. Plots changed from 3 week to 6 week cutting interval subsequently showed the same rate of production as those cut every 6 weeks throughout. Higher rates of production up to July could be explained by a higher proportion of inflorescence emergence after long growth periods. Anslow (1967) also described the crop at each harvest in terms of the quantity harvested in each successive 2.5 cm zone above soil level. This showed great differences in the erectness of the swards, with a much higher proportion of the weight of the 3 week old crop in the lower 2.5 or 5.0 cm than that which occurred in the 6 week old crop. Machine cuttings were taken to a height of about 4 cm. Cutting at 5.0 cm above ground level would remove at most about 60% of a sward cut every 3 weeks and during mid-summer, only about 35 - 40%. However, plots cut at the same 5.0 cm height every 6 weeks could be depleted of about 55 - 60% of their weight in mid-summer and more than 80% in early June. Thus, when one considers the Agassiz cutting trial, the low-infrequent cutting treatment should have significantly outyielded the high-frequent cutting, yet, on an annual basis this did not occur. When one examines the seasonal distribution of yield, it can be seen that the LI cutting yielded greater D . M . during the spring and early summer than H F cutting, but that this yield advantage for L I cutting did not persist through the drier summer months. 4.2.2.2 Cutting Height Investigations into the effect of closeness of defoliation have produced conflicting evidence. Several researchers have found that cutting 86 close to ground level increased herbage yields compared to lax cutting (Robinson and Sprague, 1947; Robinson et al., 1952; Burger et al., 1958; Reid, 1959 and Reid and MacLusky, 1960). Reid (1959) worked with a perennial ryegrass/white clover sward over 3 years and found that with either 6 or 8 defoliations per year, herbage D . M . was increased by over 40% by close defoliation to within 2.5 cm of ground level compared to lax defoliation to within 7.5 cm of ground level. Conflicting evidence has also been produced which indicates that lax cutting can outyield close cutting (Robert and Hunt, 1936 and Brougham, 1959). Important to the apparent contradiction, Janti and Kramer (1956) observed that soil moisture content influenced the effect of defoliation on herbage yields which led Appadurai and Holmes (1964) to study the interaction of closeness of defoliation (for both grazing stage and silage stage) with different soil moisture treatments. In general they found that close defoliation resulted in significantly higher D . M . yields than lax defoliation but that the effect varied with moisture regime. Under the 'wet regime', increases of 20 to 40% D . M . were recorded; under the 'medium regime' increases were only significant on the silage treatment, while under the 'dry regime' there were no significant differences and a consistent tendency at the silage stage for lax cutting to yield more than close cutting. Thus moisture regime affects the yield response to defoliation severity. From the point of view of annual yield one would expect the LI cutting to yield more than H F cutting due to the lower cutting height (as well as the longer regrowth interval already discussed). When one examines the seasonal distribution of yield this is certainly the case in the spring and early summer, but the yield under H F cutting is in general 87 greater during the dry summer period, most probably because of the improved yield obtained by lax cutting under moisture stress. 4.2.3 Seasonal Dis tr ibut ion of Y i e l d Production curves for each cultivar over all harvest seasons are presented in Figures 14 to 23. Dry matter yields at each harvest are presented in Tables 47 to 51 in Appendix I. Each curve of production was computed by obtaining average D . M . yields for each harvest interval as follows: Average Yield = Harvest D . M . Yield (Kg.) Regrowth Period (days) and plotting this at the mid-point of the regrowth interval. Linearity of growth is assumed in this method and although regrowth may indeed be non-linear (Brougham, 1959) the method is used frequently and provides production curves allowing comparisons to be made for cultivars and treatment differences. Anslow and Green (1967) assumed linear growth between harvests but used three out-of-phase sets of plots to generate production curves. Annual production showed little variability between the out-of-phase sets, but it was felt that conclusions about rates of production, from one set of plots only, would depend to a great extent on the sequence of dates chosen. By using three series it is possible to record growth rales that are typical of a particular system of herbage production. Due to resource constraints the information generated in the investigation reported here is from one series of plots. Any further investigations to generate seasonal production curves would benefit greatly by using the 88 method of Anslow and Green (1967) even though the amount of sampling and the plot area required would be greater. 4.2.3.1 Beginning of Spring Production Observations were made on a weekly basis beginning in early March in order to assess the onset of measurable spring production. The swards were visually active in late February/early March but visual increases in production were not evident at these limes. Production increases occurred generally in early/mid-March in 1984, 1985 and 1986. However, growth was much slower in 1985 and 1986 due to an extremely wet April in 1985 and a cooler than normal Apri l and wetter May in 1986. Although cultivar differences occurred in the rate of early spring production, no noticeable differences were evident as to the onset of spring growth when assessed visually on a weekly basis. Use of the weighted disc meter would greatly improve recording of the onset of spring production and give quantitative estimates of the rate of herbage production in early spring. Any further investigations in this area would benefit from the use of disc meters or height measurements (e.g. with the 'sward stick'). In general the orchardgrasses showed greater early spring production than the perennial ryegrasses for both cutting treatments in all harvest years, the only exception being the relatively high production of Barlano (44 kg D.M./ha/day) compared to Prairial (45 kg D.M./ha/day) and Sumas (34 kg D.M./ha/day) in mid -Apr i l , 1985. In general, first cut harvests were greater for orchardgrass than perennial ryegrass. As mentioned, early spring growth rates were generally higher in 1984 than 1985 and 1986 due to wetter springs in the latter two years. First cut harvests and therefore early spring growth rates in 1984 were 89 higher for LI cutting than H F cutting due to the fact that physically more herbage was harvested. Table 11 shows the first harvest in May 1984 to be greater under LI cutting. 4.2.3.2 Spring Production In all three years both Norlea and Barlano experienced rapid increases in the rate of production from mid-Apr i l to mid-May before reaching peak production. There were no differences between cutting treatments. Such rapid spring growth is typical of ryegrasses in contrast to the slower increase in spring production measured for both orchardgrass cultivars. Prairial and Sumas showed production increases from mid/late A p r i l to early July in 1984 and from mid-Apri l to early June in 1985. Production increases for both orchardgrasses were very similar to ryegrass production in 1986, increasing from mid-Apr i l to mid-May. The rapid spring production of perennial ryegrass is also evident in the individual harvest yield data where in general orchardgrass yielded significantly more herbage for first harvests and perennial ryegrass yielded significantly more herbage on second harvests, for LI cutting, and second and third harvests for H F cutting (Tables 47 to 51, Appendix I). Peak production varied with season, cultivar and cutting treatment. Orchardgrass reached peak production later than perennial ryegrass. Under L I cutting in 1984, orchardgrass reached peak production in mid-July (81 kg D.M./ha/day for Prairial and 75 kg D.M./ha/day for Sumas) while ryegrass reached peak production in late M a y (89 kg D.M./ha/day for Barlano and 100 kg D.M. /ha /day for Norlea). With H F cutting, orchardgrass reached peak production in late June (93 kg D.M./ha/day for Prairial and 102 kg D.M./ha/day for Sumas). Peak productions in 1985 90 were greatly reduced. Prairial showed maximum production under LI cutting in mid-June (61 kg D.M./ha/day) while Sumas peaked in late May at 48 kg D.M./ha/day. Both Barlano and Norlea also peaked in late May at 54 and 60 kg D.M./ha/day respectively. Peak production for H F cutting occurred in early June for Prairial, Barlano and Norlea (56, 62 and 59 kg D.M. /ha /day respectively) while Sumas reached maximum production in late May (50 kg D.M./ha/day). During 1986 all four cultivars reached peak production during late May at levels higher than the previous two years (103 for Prairial, 127 for Sumas, 122 for Barlano and 130 kg D.M./ha/day for Norlea). Orchardgrass generally has a slower increase in spring production and a later date of peak production which enables herbage to be produced at greater levels in late and mid-summer than is possible for perennial ryegrasses which produce a greater portion of their yield during the spring/early summer period. However, this yield distribution can be significantly affected by cutting treatment. Under LI cutting in 1984, orchardgrass produced 34.5% of annual production in the spring ( A p r i l / M a y ) and perennial ryegrass produced 53.5% of its annual production, whereas, under HI cutting a portion of annual yield was shifted to the summer (Table 11). Orchardgrass produced 21.2% and perennial ryegrass 28.1% under HI cutting in the spring of 1984. Similar results occurred in 1985 with orchardgrass producing 46.2% and perennial ryegrass 58.3% of annual production in the spring under LI cutting compared to 36.3% for orchardgrass and 34.9% for ryegrass under H F cutting. The proportionately lower yield in 1985 compared to 1984 under H F cutting can be attributed to the very dry summer and greatly reduced herbage growth in 1985. The much greater differences in spring production between L l and H F cutting are more in agreement with other studies than was previously indicated by only examining annual yield data. LI cutting yielded more than H F cutting during the spring when moisture was not limiting. A major reason for this could be that a greater portion of a sward under LI cutting is present at a higher level above the ground due to the longer regrowth period and the fact that physically more herbage is harvested by the lower cutting height. First cut harvests under LI cutting yielded higher because more herbage was physically removed. If this first cut yield advantage is removed, the annual yields were virtually the same for all cultivars over both cutting treatments. First cut yields in 1986 were generally lower than in previous years. Although only the LI cutting was continued through 1986 both LI and H F cut plots were harvested on May 7, 1986 to the same low cutting height. Yields are presented in Table 12. For Prairial, Barlano and Norlea, LI yields were higher while the reverse was true for Sumas with higher H F yields. A possible explanation is that the cutting treatments produced morphological differences with low cutting producing a lower, denser sward with more tillering and leaf production than higher cut swards. Appadurai and Holmes (1964) worked with ryegrass/white clover swards and found that increased leaf production (particularly early in the season) was a major factor resulting in higher yields obtained from close defoliation. A further factor involved was that Net Assimilation Rate was also greater after close defoliation (Brougham, 1956a). The lower yield for Sumas under LI cutting could be attributable to a cutting effect on that particular cultivar. Sumas generally produced lower first cut harvests than Prairial. First cut harvests for Sumas in 1985 were very similar for LI (1443 kg DM/ha /day) and H F (1450 kg DM/ha /day) treatments even though cut at different heights. In 1985 H F yields would also have been TABLE 11: SEASONAL PRODUCTION AT AGASSIZ. 1984 TO 1986 CUTTING CULTIVAR SEASON TREATMENT Spring (April/May) Summer (June to Aug) Fall (Sept/Oct) Yield percent Yield percent Yield percent 1984 Ll Prairial Sumas Barlano Norlea HF Prairial Sumas Barlano Norlea 4556 35.8 3749 33.1 4285 44.5 4574 62.4 2501 21.0 2336 21.3 2412 25.9 2335 30.2 (kg DM/ha) 6501 51.0 6026 53.3 3470 36.1 1055 14.4 7656 64.2 7265 66.4 5770 62.0 4966 64.2 1682 13.2 1545 13.6 1868 19.4 1700 23.2 1767 14.8 1346 12.3 1128 12.1 436 5.6 1985 Ll Prairial Sumas Barlano Norlea HF Prairial Sumas Barlano Norlea 3340 46.3 2976 46.0 3434 56.5 2511 60.0 2264 33.0 2498 39.6 2238 36.9 1372 32.9 2897 40.2 2618 40.5 1987 32.7 1229 29.4 3421 49.9 3231 51.2 2628 43.4 2354 56.5 973 13.5 872 13.5 653 10.8 446 10.6 1172 17.1 577 9.2 1195 19.7 444 10.6 1986 LF- Prairial 4553 38.0 5844 48.8 1589 13.2 Sumas 4662 43.1 4836 44.7 1327 12.2 Barlano 4475 46.4 3752 38.9 1422 14.7 Norlea 3865 45.6 3554 41.8 1068 12.6 to 93 higher than LI yields if both were cut low. Thus, the low cutting height slows Sumas growth in the early spring, a result that could be due to several factors including leaf area, root reserves and growth morphology of Sumas. T A B L E 12: Yield at First Harvest. Agassiz. 1986. Cul t ivar Cutting Yie ld at first harvest T r e a t m e n t (kg D.M./ha) P r a i r i a l L I 1401 a Pra ir ia l HF 1256 ab S u m a s H F 1169 be Sumas L I 945 cd Barlano L I 781 de Bar lano H F 532 ef Nor lea L I 514 f Nor lea H F 244 g M e a n 855 S.E. of Mean 117 L . S . D at P = 0.05 252 Note : Means within the same column with different letters are significantly different at P = 0.05. 94 4.2.3.3 Summer Production Throughout the three years all four cultivars suffered decreases in production during the dry summer periods which varied from year to year. Generally, reduction of growth rates in orchardgrass was not as severe as those experienced in ryegrass. Production declined steadily for orchardgrass for the rest of the season from peak yield in 1984 (LI and H F cutting), 1985 (only LI cutting) and 1986. By late August growth rates were still in excess of 40 kg D.M./ha/day during 1984 and 1986. During the very dry summer of 1985 orchardgrass production dropped to below 20 kg D.M./ha/day and only under H F cutting did production increase in response to fall rainfall. Ryegrass suffered decreasing summer production that varied with cutting treatment. Under L l cutting both ryegrass cultivars showed the 'traditional' grass growth curve with maximum spring production, decreased summer production followed by an increase in fall growth rales. In 1984 and 1985 Norlea essentially stopped growth during late July to mid-August, with growth reduction being extremely severe in 1985. Barlano had reduced yield in 1984 (38 kg D.M./ha/day) and also suffered from moisture stress and declining growth in 1985. Both ryegrasses experienced a different growth pattern under H F cutting. In 1984 Barlano and Norlea had higher growth rates from late July to late August (45 - 65 kg D.M./ha/day for Barlano and 30 - 46 kg D.M./ha/day for Norlea) than under L l cutting. During 1985 Norlea growth was improved under H F cutting. Barlano growth declined to zero under H F culling but improved during August to yield more herbage than LI cutting. Production during 1986 under LI cutting showed a steady decline in growth rates to about 18 kg D.M./ha/day in late August. 95 Herbage yields during the summer were significantly affected by cutting treatments. H F cutting increased summer yields when compared to LI cutting by shifting production from the spring (Table 11). During 1984, LI cutting resulted in 51.7% of ryegrass annual yield being produced in the summer compared to 65.3% of orchardgrasses and 63.1% of ryegrass annual yield under H F cutting. Similarly in 1985 LI cutting resulted in 40.3% of orchardgrass and 31.1% of ryegrass annual yield being produced in the summer compared to 50.6% of orchardgrass and 50.0% of ryegrass growth under H F cutting. These results agree with other studies which indicate that production is increased under lax cutting when the sward is moisture stressed (Appadurai and Holmes, 1964). With the low cutting of the L I treatment combined with moisture stress, growth rates were more depressed and the herbage took longer to recover when compared to a more lax cutting height. It can be seen that a more even distribution of yield can be obtained during late spring/early summer and late summer/mid-summer. A more even distribution of yield over these periods has obvious advantages, particularly under grazing, and may be achieved by using different species (i.e. orchardgrass vs perennial ryegrass) in separate swards or possibly mixed swards and by variation in defoliation intensity and frequency. Under LI cutting in 1986, similar distributions of annual yield are evident. However, the Norlea plots did show comparatively less yield in the spring and more yield in the summer when compared to 1984 and 1985 probably because of its greater orchardgrass content (18%) and because much of the summer yield was produced during June, rather than the drier July and August. 96 4.2.3.4 Fall Growth There were no consistent cultivar cutting treatment trends in fall (September/October) production. In most cases production declined steadily from late summer through the fall. Increases in fall production were seen in ryegrass under L I cutting in 1984 and L I and H F cutting in 1985. Prairial and Sumas also showed increased fall production under H F cutting in 1985. A l l of these resulted from increased production after rainfall following the dry summer period of reduced growth. Figures 14 to 23 show similar levels of fall growth (15 - 25 kg D.M./ha/day) but the 'fall flush' is not evident because summer growth rates were maintained at higher levels (eg. ryegrass under H F cutting in 1984). Orchargrass produced about the same relative yield during the fall period irrespective of season and cutting treatment (13.3% of annual yield). Ryegrass showed more variability with Norlea ranging from 5.6% under H F cutting to 23.2% under LI cutting in 1984. On average ryegrass produced 13% of annual yield in the fall. 4.2.3.5 Length of the Growing Season After the last harvest of the year, plots were observed for continued growth. No cultivar or cutting differences were evident in the duration of growth through the fall and into the winter. In 1984 growth was evident until late November although not enough material was produced for another harvest. The dry summer in 1985 seemed to have a residual effect on growth with fall yields being quite low. Growth under LI cutting continued after the last harvest on September 20 to late October, but no harvestable growth was produced. After the last H F harvest on October 25 growth also stopped. Early frosts occurred in the fall of 1985 followed by 9 7 FIGURE 14 : GROWTH CURVE FOR PRAIRIAL AND SUMAS - AGASSIZ 1984. >» ro T3 To x: Q cn HI I— < CC X I-£ o tc o 140 120 100 -80 -60 40 20 LOW-INFREQUENT CUTTING PRAIRIAL SUMAS APL MAY JN J L AUG SEP O C T MONTH 9 8 FIGURE 16 : GROWTH CURVE FOR PRAIRIAL AND SUMAS - AGASSIZ 1984. ro TO ro JC O) LU < DC O DC a 140 120 100 80 60 40 20 APL MAY JN HIGH - FREQUENT CUTTING PRAIRIAL SUMAS JL MONTH AUG SEP 0C^ FIGURE 17 ; GROWTH CURVE FOR BARLANO AND NORLEA - AGASSIZ 1984. ro -o ro 140 120 100 HIGH - FREQUENT CUTTING • BARLANO • N O R L E A JL MONTH OCT FIGURE 18 : GROWTH CURVE FOR PRAIRIAL AND SUMAS - AGASSIZ 1985 FIGURE 19 ; GROWTH CURVE FOR BARLANO AND NORLEA - AGASSIZ 1985. OCT MONTH 101 FIGURE 22 ; GROWTH CURVE FOR PRAIRIAL AND SUMAS - AGASSIZ 1986. _ 140 « ro x: 2 Q CD 111 < CC X I— o CC o 120 100 F 80 60 40 20 • LOW-INFREQUENT CUTTING • PRAIRIAL • SUMAS If - II — • — • - * APL MAY JN J L MONTH AUG SEP OCT FIGURE 23 : GROWTH CURVE FOR BARLANO AND NORLEA - AGASSIZ 1986. o o I 1 APL MAY JN J L AUG SEP OCT MONTH 102 a colder than normal November. In 1986 growth was evident until early November but, again, no harvestable yield was produced. 4.3 A B B O T S F O R D C U T T I N G T R I A L 4.3.1 Annu.nl Yie lds Annual yields are shown in Table 13 and Figure 24 for all four cutting treatment combinations. The LI treatment significantly (P<0.05) outyielded the other three treatments. The treatments ranked as follows: LI > L F > HI > H F . There were no significant differences between the last three treatment y i e l d s . A v e r a g e annual y i e lds for the Frequent/Infrequent and H i g h / L o w cutting treatments are reported in Table 14. L o w cutting (2.5 to 4 cm) significantly (P<0.()5) outyielded high cutting (7.5 to 10 cm) while infrequent cutting (3 harvests) outyielded frequent cutting (4 harvests) although the difference was not significant (P>0.05). These results arc in general agreement with several trials which have examined the effect of both cutting frequency and cutting intensity on annual production as previously discussed. 4.3.2 Seasonal Yie lds : Under all four cutting treatments the major portion of annual yield was obtained from the first harvest at conservation stage (hay and/or silage) made on June 13. The yield at each harvest for frequent cutting is shown in Figure 25 and reported in Table 15. Three harvests were made after the cut. Yie ld was especially depressed in the midsummer dry period and recovered somewhat in the fall . The yield for the two post-conservation cuts under infrequent cutting is shown in Figure 26 and 103 reported in Table 15. First harvest yields showed no significant (P>0.05) differences across all four treatments (Table 16) even though the low cutting treatments were cut 5 to 6 cm lower than the high cutting treatments suggesting that the proportion of yield contained in the horizon from 0 to 2.5 - 4 cm was not significant. This would agree with the growth stage of the herbage, which was mostly in boot and heading, where herbage yield in the 0 to 2.5-4 cm horizon would consist mostly of stems. Subsequent harvests showed the same general trend as annual yields with low cutting yielding more than high cutting and infrequent cutting yielding more than frequent cutting. The total post-conservation harvests showed the same general trend in treatment yields (Table 16). A l l treatments had their last harvest taken on October 15. In contrast to the first harvest, low cutting yielded significantly more herbage than high cutting. In this case the herbage had a different growth form, namely vegetative, where a much greater proportion of the herbage was contained in the 0 to 2.5-4 cm horizon resulting in more herbage actually being harvested by low cutting. It is also possible that some of the increased yield could have resulted from low cutting increasing sward density and tillering, however, if this did increase yield it is probably of small significance since the effect was probably not that great from only 4 months of harvesting. In contrast to the Agassiz trial there was no apparent yield advantage during the dry summer months as a result of H F or even HI cutting when compared to both low cutting treatments. The rapidly draining soils combined with the extremely dry weather resulted in growth stopping for all cutting treatments for about three weeks. Visual observation revealed that growth was maintained for a longer period T A B L E 13: A N N U A L YIELD FOR A B B O T S F O R D CUTTING TRIAL. 1985 CUTTING T R E A T M E N T ANNUAL YIELD (kg D.M. / ha) LOW - INFREQUENT 8,207 a LOW - F R E Q U E N T 6,721 b HIGH - INFREQUENT 6,066 b HIGH - F R E Q U E N T 5,674 b MEAN 6,667 S . E . O F MEAN 465 L.S.D. 1,053 Note: Means with different letters are significantly different ( P=0.05) T A B L E 14: A N N U A L CUTTING T R E A T M E N T YIELDS. A B B O T S F O R D 1985 CUTTING T R E A T M E N T ANNUAL YIELD (kg D.M. / ha) 104 F R E Q U E N T INFREQUENT 6,197 a 7,137 a HIGH LOW 5,870 b 7,464 a MEAN S . E . O F MEAN L.S.D. 6,667 465 1,053 Note: Means with different letters are significantly different ( P=0.05) T A B L E 15: INDIVIDUAL H A R V E S T YIELDS FOR A B B O T S F O R D CUTTING TRIAL. 1985 CUTTING HARVEST DATE T O T A L T R E A T M E N T JUNE 13 JULY 12 AUG 28 OCT 15 YIELD Yield (ka D.M. / ha) LOW-FREQUENT 4,189 a 1,045 a 200 a 1,287 a 6,721 a HIGH-FREQUENT 4,197 a 652 b 109 a 716 b 5,674 a M E A N 4,193 849 155 1,002 6,198 S . E . O F M E A N 362 113 40 137 376 L.S.D. 1152 361 127 437 1195 CUTTING H A R V E S T DATE T O T A L T R E A T M E N T JUNE 13 JULY 25 O C T 15 YIELD Yield (kg D.M. / ha) LOW-INFREQUENT 4,595 a 1,930 a 1,682 a 8,207 a HIGH-INFREQUENT 4,055 a 1,174 a 837 b 6,066 b M E A N 4,325 1,552 1,260 7,137 S . E . O F M E A N 510 318 235 595 L.S.D. 1624 1012 750 1893 Note: Means with different letters within the same column are significantly different (P=0.05). TABLE 16: YIELD AT FIRST AND LAST HARVESTS ABBOTSFORD CUTTING TRIAL. 1985 CUTTING FIRST LAST YIELD TREATMENT HARVEST HARVEST EXCLUDING JUNE 13 OCTOBER 15 FIRST HARVEST LOW - INFREQUENT 4,595 a 1,682 a 3612 a LOW - FREQUENT 4,189 a 1,287 a 2532 b HIGH - INFREQUENT 4,055 a 837 b 2011 be HIGH - FREQUENT 4,197 a 716 b 1477 c MEAN 4,259 1,131 2408 S.E. OF MEAN 386 189 292 L.S.D. 873 426 659 Note: Values with different letters within the same column are significantly different ( P=0.05 ). o FIGURE 24: ANNUAL YIELD AT ABBOTSFORD CUTTING TRIAL, 1985 YIELD (kg D.M./ha) (Thousands) H I G H L O W CUTTING HEIGHT F I G U R E 2 5 : Y I E L D AT E A C H H A R V E S T A B B O T S F O R D , F R E Q U E N T C U T T I N G , 1 9 8 5 YIELD (kg D.M, / ha) (Thousands) JUNE 13 JULY 12 AUG 28 OCT 15 H A R V E S T DATE FIGURE 26: YIELD AT EACH HARVEST ABBOTSFORD, INFREQUENT CUTTING, 1985 YIELD (kg D.M. / ha) (Thousands) JUNE 13 JULY 25 OCT 15 HARVEST DATE CUTTING TREATMENT INFREQUENT - HIGH INFREQUENT - LOW 1 10 under III cutting than under LI cutting but the growth rate was extremely low. The last harvest under infrequent cutting was made on October 15 from growth that had accumulated since mid-July. Most of the yield under LI cutting had accumulated in the latter part of the growth period. Whether there were any actual yield differences between LI and HI cutting during the dry period cannot be determined since no harvests were made at that time. The August 28 harvest under frequent cutting showed a tendency toward a greater yield from low cutting compared to high cutting, although the difference was not significant (P>0.05). Any yield advantage that may have resulted from high cutting did not occur, possibly because of the complete lack of growth under all treatments during the dry summer. 4.4 O Y S T E R R I V E R F E R T I L I Z E R T R I A L 4.4.1 A n n u a l Yie lds : Annual yields are shown in Table 17 and Figure 27. Application of 300kg N/ha/year (cutting regimes II, III and IV) yielded significantly more herbage than an application of 100kg N/ha/year (cutting Regime 1). A single area of the same pasture was sampled on the same dates without any nitrogen application. Annual yield on this area was 7,080 kg D.M./ha/year and is not included- in the statistical analysis because no replicate samples were taken. Increasing D . M . yield in response to nitrogen application has been demonstrated in numerous experiments (Cowling and Lockyer, 1965; Brockman, 1966; Reid, 1966; Kallofen et al., 1966; Wolfe and Crofts, 1969; Wilman and Wright, 1981; Reid, 1985 and Large et al., 1985). The average yield of 10,228 kg D . M . / h a is comparable 111 to yields in the same area, at Agassiz and even with trials in the U . K . (Reid, 1985). The yield of the unfertilized area (7,080 kg D.M./ha) and Regime I (7,706 kg D.M./ha) are higher than what would be expected from a pure grass sward at 0 and 100 kg N/ha respectively. The relatively higher yield can be attributed to contribution of nitrogen from the clover portion of these plots. Clover content was significantly (P<0.05) higher on the 0 N (35%) and 100 N (29%) plots compared to the 300 N (mean=8%) plots. The contribution of N from clover can be quite significant and tends to vary directly with the percent clover content in the sward. Cowl ing and Lockyer (1965) demonstrated a fertilizer nitrogen equivalent from white clover of 162 to 191 kg N/ha for perennial ryegrass and 138 kg N/ha for orchardgrass during a three year trial. It is not unreasonable to expect such magnitudes of fertilizer nitrogen to be contributed by the clover portion in this trial. Clover N yields of up to 250 kg/ha have been recorded . Since no pure grass sward at 0 N was examined in this trial it is not possible to estimate the percent recovery of N or the D . M . response to applied N in comparison to a pure grass sward. The relative D . M . response of regimes II, III and IV when compared to I were 12.7, 12.7 and 12.4 kg D . M . / kg N respectively and when compared to the 0 N plot, 10.6, 10.6 and 10.4 kg D . M . / kg N respectively. The relative D . M . increase of regime I when compared to the 0 N treatment was 6.3 kg D . M . / kg N . These responses are within the range reported by several experiments with grass clover swards. Chestnutt and Lowe (1970) summarized responses mainly from cutting trials in the U .K.and gave a range of 3.7 to 17.4 kg D .M. / kg applied N (mean of 8.8) from several N rates up to 390 kg/ha. Frame and Boyd (1987a) examined several trials and showed a range of responses TABLE 17: DRY MATTER YIELD FOR OYSTER RIVER FERTILIZATION TRIAL FERTILIZATION HARVEST DATE TOTAL REGIME MAY 9 JUNE 10 JULY 8 AUG 12 SEPT 11 OCT 29 YIELD Yield (kg D.M./ha) 1985 I 2,561 a 3,307 b 500 c 887 b 202 c 249 c 7,706 b II 2,753 a 4,056 a 902 a 1,695 a 357 b 485 b 10,248 a III 2,552 a 3,896 a 723 ab 2,058 a 472 a 544 b 10,245 a IV 2,741 a 4,153 a 643 be 1,230 b 321 b 1,104 a 10,192 a MEAN YIELD 2,652 3,853 692 1,467 338 595 9,597 S.E. OF MEAN 236 222 80 173 39 50 310 LS.D. 534 503 182 391 89 113 702 Note: Different letters within the same column are significantly different ( P = 0.05 ). F I G U R E 2 7 : ANNUAL Y I E L D AT O Y S T E R R I V E R YIELD (kg D.M./ha) FERTILIZATION REGIME 114 from -6 to 20 kg D . M . / k g applied N for rates of 0 to 400 kg N/ha. The variation in response can be considerable and will vary from year to year and location to location. Response of regimes I, II and III would be considered above levels of marginal profitability in Europe where 7.5 (Prins et al. , 1980) and 10 kg D . M . / k g applied N (Morrison, 1980) were target values. Such values will obviously be different in B . C . depending on feed and fertilizer costs. 4.4.2 Seasonal D i s t r i b u t i o n of Y i e l d : 4.4.2.1 Spring Production: Figure 28 shows the seasonal distribution of yield for the four cutting regimes. The method used to generate the curves was the same as that used to generate the Agassiz growth curves. Growth increased rapidly from late March to peak production in late May. Rates of production for fertilizer regimes II, III and IV were very similar throughout the spring while the growth rate of I increased at a slower rate from mid-Apri l . Peak production was very similar for regimes II, III and IV being 127, 122 and 130 kg D.M./ha/day respectively. Regime I had a somewhat lower peak production at 103 kg D.M./ha/day. Harvest results in Table 17 show that first harvest yields on M a y 9 were not significantly (P>0.05) different whereas yields for regime I were significantly (P<0.05) lower on June 10 when compared to the other three regimes. Even though regimes II, III and IV received different levels of N fertilizer on May 9 (125, 100 and 50 kg N/ha respectively), their yields on June 10 were not significantly (P>0.05) different. Yie ld responses relative to regime I were 6.0 kg D . M . / kg N for II and 5.9 kg D . M . / kg N for HI whereas IV showed more 1 1 5 efficient production with a yield response of 16.9 kg D . M . / kg N . For the four week growth interval the application of 50 kg N/ha proved as effective as 100 and 125 kg N/ha for production at the next harvest. Such differences could be attributable to the way in which a sward responds to different levels of applied N and how measured growth response is determined by harvest interval. Wilman (1965), working with Italian ryegrass, showed that although total N content of herbage showed differences at the end of the first week after applications of 28 and 84 kg N/ha there were no differences in yield response until the end of the second or third week. The effect was even more pronounced at levels of 84 and 140 kg N/ha where differences in herbage N content were apparent at the end of the second week, but yield differences did not occur until after the fourth week. If such effects occurred at Oyster River the yield advantages of higher N applications would be apparent beyond the 4 week growth interval and appear in later harvests. This certainly appeared to be the case since increased yields for II and III were evident for the August 12, September 11 and October 29 harvests even though the last applications of N were made on June 10. The same trend is also evident for IV where an N application on June 10 gave good D . M . yields on August 12 even though no N was applied on July 8. In general the response to an N application seemed to occur at the second harvest after application. 4.4.2.2 Summer Production: Growth rales dropped markedly throughout June to reach lows from 18 to 32 kg D.M./ha/day. This drop is partly attributable to the normal decline in production as mid-summer approaches and partly clue to the 116 lack of adequate rainfall for continued growth. Yields on July 8 were quite low (500 to 902 kg D.M./ha) with II and III yielding significantly (P<0.05) more herbage in response to previous N applications. Production showed an increase for all regimes during July as a result of 3 weeks of irrigation from late July to early August. Irrigation was applied as part of farm management and no measurements of the amount of water applied were made. Significant responses occurred after irrigation with yield and growth rates in the following order: III > II > IV > I. A l l four treatments reached a peak in production during late July with growth rates of 59 (for III), 48 (for II), 35 (for IV) and 25 kg D.M./ha/day for I. Yie ld responses to fertilizer were especially evident for regimes II and III as a result of June 10 and possibly May 9 N applications. 4.4.2.3 Fall Production: Irrigation was not continued in August and as a result production decreased due to moisture stress. September 11 harvests were low ( 202 to 472 kg D.M./ha) but regime III produced significantly (P < 0.05) more herbage in response to the July 8 application. As with the drop of production in June / early July, the lack of moisture greatly reduced yield response to applied N . Growth rates declined slowly for regimes I, II and III as fall approached. However, regime IV experienced an increase in production (23 kg D.M./ha/day) in response to the August 12 application of 100 kg N/ha. Yie ld on October 29 was significantly higher for IV. Yields of II and III were greater than I, probably due to some residual N effect on improved root reserves and plant vigour developed by earlier applications. 1 17 Throughout the season it can be seen that seasonal distribution of growth was altered by the various split applications of N when moisture was not limiting. Several studies have investigated the effect of distribution of yield from single and split applications with the same conclusions. In general split applications produce a more even yield distribution without affecting annual yields when compared to single applications (Kaltofen et al., 1966). Under local conditions Bomke and Bertrand (1979) noted that three applications shifted yield from the spring to summer when compared to a single spring application. This has obvious advantages for hay production in the South Coastal region of B . C . and under grazing conditions increased summer production can be very valuable. The Oyster River trial demonstrated that various patterns of split application can vary yield distribution without affecting annual yield. The only disadvantage of split applications is the need for more labour and equipment time. Large, single, applications in spring are relatively easy to apply and readily contracted out. It has also been noted that large, single, N applications can even out seasonal yield more than smaller single applications. Kaltofen et al. (1966) found that split applications are of less importance in evening out distribution of yield at high N rates (320 to 480 kg N/ha) than at lower N rales (80 to 160 kg N/ha). High, single applications in the spring had a lasting effect with yields of the fourth utilization markedly increased. Even when N is applied in even portions over the season, large total applications produce more even distribution than smaller applications (Morrison, 1980). Cowling and Lockyer (1965) applied 0 to 390 kg N/ha in even applications throughout the season and noted that higher applications produced a greater portion of production during the summer months. FIGURE 28 : GROWTH CURVES FOR OYSTER RIVER - 1985 140 OCT MONTH F I G U R E 2 9 : Y I E L D AT E A C H H A R V E S T O Y S T E R R I V E R , 1 9 8 5 YIELD (kg D.M./ha) (Thousands) 5 i MAY 9 JUNE 10 JULY 8 AUG 12 SEPT 11 OCT 29 HARVEST DATE 120 4 . 4 . 3 Herbage Nitrate Levels: Problems of high herbage nitrate levels and surface water pollution must be considered when large single applications of N are applied. Wright and Davison (1964) considered forages containing more than 0.34 to 0.45% nitrate N should be regarded as potentially lethal. Muhrer et al. (1956) suggested that a level of 0.07% nitrate N is critical. Cattle were fed forage containing sub-lethal levels of nitrate and subsequently had lactation and reproduction problems. ap Griffith (1960) placed the lower level of nitrate toxicity at 0.22%. Reid (1966) demonstrated that annual applications greater than 334 kg N/ha (applied in five equal applications) produced herbage nitrate levels in excess of 0.22%. Applications in excess of 400 to 500 kg N/ha greatly increased nitrate content (Deinum and Sibma, 1980). They also demonstrated that herbage nitrate content can vary considerably with cutting management. Even applications to a total of 600 kg N/ha/year resulted in much lower nitrate levels under a system of one week regrowth (30 cuts; N applied at 20 kg/ha/cut) than compared to a seven week regrowth system (4 cuts; N applied at 150 kg/ha/cut). Therefore even applications at fairly high levels can result in less nitrate with frequent defoliation. This was attributed to the fact that more frequent defoliations resulted in a leafy, young sward with low nitrate levels in the leaves compared to an older sward with higher nitrate levels in the stems and old tissues. It was also noted that fitting N application to seasonal yield (for example, heavy spring applications) resulted in higher nitrate in the spring and summer when more stemmy crops were harvested. The combination of defoliation with N application has also been 121 shown to increase nitrate levels in regrowth (Alberda, 1960). It was speculated that defoliation decreases root and stubble carbohydrates which results in less nitrate reduction in the roots and stubble and a consequent translocation of nitrates to the leaves (Deinum and Sibma, 1980). In view of nitrate content alone, individual applications of N should not normally be large. 4 . 4 . 4 Ni trogen Appl icat ions and Regrowth Per iod: The argument for smaller individual N applications is also supported by the fact that growth response to N is affected by regrowth periods. Brockman (1966) noted that N uptake by grass is more rapid than growth response. Thus, for a short regrowth period, which would be especially common under grazing situations, there is an incomplete response to heavy N applications followed by a residual effect after harvesting. This effect was noted at Oyster River with the general growth response to higher N occurring at the second harvest after application. Response to smaller applications is seen mostly in the first harvest after application with little residual effect (Kaltofen et al . , 1966 and Brockman, 1966). Under relatively short regrowth periods with grazing ( two to four weeks) large N applications could be wasteful since total response will rely on residual effects. In wetter climates the residual effect may be lost to N leaching from the soil. During one year of their trials Kaltofen et al. (1966) noted a significant yield advantage from split N applications compared to single spring applications attributable to N leaching as a result of heavy spring rains. This is an important factor to consider in South Coastal B . C . during spring. Dowdell et al. (1980) noted as much as 52% leaching losses (mostly over the winter) at high N rates (750 kg N/ha/year) whereas at lower 122 rates (250 kg N/ha/year) losses were negligible. Smaller applications after defoliation may be more effective than relying on residual effects. Wi th silage and hay systems the grass has longer regrowth periods in order to respond to high applicat ions. However , even considering the longer g rowing period, B rockman (1966) recommended that applications be smal l and frequent for both grazing and silage systems. In silage systems uptake of recoverable N was very high over the growth period resulting in little residual effects after the first harvest and leading to poor overal l response through the year. M o r e recent work by Prins et a l . (1980) has explained in more detail the residual effects of applied N . They conducted a series of seven field experiments over four years to demonstrate grass response to N during the grazing season in relation to different levels and intensities. A system of 80 kg N/ha/cut gave a greater residual effect than an application of 60 and l i k e w i s e , an intensive system of 120 resulted in even greater residual responses than 80 kg N/ha/cut. S o i l analysis revealed that residual effects can only be partly explained by increases in so i l mineral N . It was concluded that after large applications, part of the N is stored in roots and stubble to be available for regrowth. D i l z (1966) had shown similar results with perennial ryegrass. Such residual effects are evident in the seasonal yields at Oyster River . 4.4 .5 Distribution of Yield and Effect of Moisture: Regimes II, III and I V shifted y i e l d away from the spring ( A p r i l / M a y ) when compared to I (Table 18). On average II, III and IV produced 65% of their y ie ld in the spring when compared to 75% for regime I. Regimes II and III shifted this production into the summer 123 (June/July) with 25.3% and 27.2% respectively of the annual yield during this period compared to 18% for 1 and 18.4% for IV. No appreciable differences were noted for late August / early September. Regime IV shifted yield to the fall with 10.8% compared to a 4.4% average for the other three regimes. T A B L E 18 : Percent Distribution of Annual Yield at Ovster River. 1985. Fert i l izer Harvest Dates Regime May 9 June 10 July 8 Aug 12 Sept 11 Oct 29 Percent of Annual Yie ld I 33. .2 42.8 6.5 11. .5 2.8 3.2 I I 26. .9 39.6 8.8 16. .5 3.5 4.7 I I I 24. •9 38.0 7.1 20. .1 4.6 5.3 I V 26. .9 40.7 6.3 12. .1 3.2 10.8 These results are similar to those of Morrison (1980) who increased summer yields by 10% by applying more fertilizer later in the season compared to the same annual rate (300 kg/ha) with more applied earlier in the season. The distribution of yield shown in Table 18 was not as good as could be expected from the fertilizer regimes employed. Differences in yield from late June / early July and late August / early September are quite small and yields are relatively low. This can be attributed to a significant extent to the lack of moisture during these periods. Higher yields and belter distributions in late July / early August occurred as a result of 124 irrigation. With the lack of adequate soil moisture yields did not respond to N applications on June 10 and August 12 in the next harvest. Responses were seen later as evidenced by the increased yield of regime IV on October 29 in response to the August 12 N application. T o insure immediate and effective use of N it is necessary that there be sufficient water in the surface soil at the time of its application (Garwood et al., 1980). - When water is supplied, the response is variable with apparent relationships between volume applied, size of the soil water deficit before irrigation, time of defoliation and time of N application. Unless N applied during dry periods is recovered later in the season (as occurred in late July / early August and to an extent in the fall), loss by leaching can occur in the winter. Where rainfall is low, full utilization of applied N can only be attained when rainfal l is supplemented by irrigation. In certain circumstances, such as grazing situations, it may be beneficial to sacrifice N utilization and apply high rates of N in the beginning or just before dry periods to produce some feed during such limes (Brockman, 1966 and Morrison, 1980). 4.4.6 F ir s t Harvest Yie lds in 1986: A l l plots were harvested on May 6, 1986 with no additional fertilizer applications since those described for 1985. Yields are presented in Table 19 and it is clearly evident that regime IV produced significantly higher yields. Compared to IV , production from I was 60.0% and average production from II and III was 79.6%. Due to leaching of N through the winter it is not reasonable to expect this yield advantage to be attributable to soil N . Prins (1978) observed that very high rates of N application in one season affect the sward in the following season. It is most likely that 125 the effect observed at Oyster River is of a similar nature with increased production arising from N stored in the roots and stubble of herbage under regime IV over the winter from the August 12, 1985 application of 100 kg N/ha (Prins et al., 1980). In terms of overall yield, the effect observed in IV is important since greater yields can be produced by this fertilization regime. It should be noted that this data is only from one year of observations and similar experiments should be conducted to see if the effect is consistent. Table 19: First Harvest Yield at Ovster River. 1986 Fert i l izer Yield Percent Regime (kg D.M./ha) Relative to IV I 1341 c 60.0 II 1849 b 81.3 III 1770 b 77.8 I V 2274 a 100.0 M e a n 18 09 S.E. of Mean 169 L.S.D. 377 Note: Means with different letters are significantly different at P = 0.05. 4.4.7 Botanica l Compos i t i on : At the beginning of the trial, ground cover determinations indicated 17% clover (mixture of red and white), 65% grass (orchardgrass and 126 ryegrass) and 18% bare soil or litter. Throughout the growing season there were progressive changes in the composition of the swards under the different regimes (Table 20). Those swards receiving 0 or 100 kg N/ha/year showed increases in clover content to 27 and 24 % respectively with corresponding decreases in grass to 48% and 57% respectively. In contrast, the three regimes receiving 300 kg N/ha/year experienced a decrease in clover to an average of 6% and a corresponding increase in grass to an average of 71%. High (300kg N/ha/year) and Low (0 and 100kg N/ha/year) application rates resulted in significant differences in percent grass and clover. Visual observation of the 0 and 100 N plots indicated that there was more red than white clover (about 2:1). These results are typical of the effect of N application on grass/clover swards (Frame and Boyd , 1986a, 1987a, 1987b). Levels of 300kg N/ha are not conducive to maintaining high clover content in swards. Not surprisingly, increasing levels of N result in greater clover reductions (Laidlaw, 1980; 1984; Laissus, 1983 and Morrison et al, 1983). Spring applications suppress clover to a greater extent than equal levels of N applied in the fall (25-75 kg N/ha). With both spring and fall applications (125 kg N/ha/year), clover is suppressed to the greatest extent when higher proportions of N arc applied in the spring (Frame and Boyd, 1986b). In general, fertilizer N increases total herbage production but decreases white clover content and production.. It is necessary to evaluate the increased production from fertilizer use against the negative effect on white clover performance. It is interesting to note that Reid (1985) demonstrated virtually the same milk yields between two groups of cows grazing a pure perennial ryegrass pasture (receiving 360 kg N/ha/year in 5 equal dressings) and a perennial ryegrass - white clover pasture (receiving 180 127 kg N/ha/ycar in 5 equal dressings). It seems that the value of white clover exists and that its benefits have been underemphasized and its problems overemphasized. Content of white clover in swards depends on many soi l , cl imatic and management factors. The maintenance of beneficial levels of clover in swards receiving N requires careful management and may be beyond management control under certain climatic and soil conditions. Relatively small N applications in the spring and fall (75 kg/ha) are too high for clover persistency (Frame and Boyd, 1986a) and allegedly N-tolerant large leaved clover varieties seem not to be more N-tolerant than smaller leaved varieties but have been shown to be more productive at all N rates, including no N (Frame and Boyd, 1987b). Under high intensity dairy farming (400 kg N/ha/year), the maintenance Table 20: Percent Ground Cover at Oyster River. September 1985 Fert i l i zer Grass Clover Weeds Litter and Regime Bare Ground Percent Ground Cover I 57 be 24 a 1 a 18 a II 73 a 7 b 0 a 20 a III 68 ab 5 b 2 a 25 a I V 73 a 7 b 0 a 20 a 0 N 48 c 27 a 2 a 23 a M e a n 64 1 4 1 22 S E of Mean 6 4.5 not sig. not sig. L.S.D. 1 3 10 not sig. not sig. Note: Means within the same column with different letters arc significantly different (P=0.05). 128 of clover yields is not a management objective. More extensive operators like beef and sheep who traditionally apply less N may benefit from clover-grass swards. However, management of such mixed swards, with applied N , is difficult and recent suggestions of high clover-grass swards with no N and pure grass swards with N application combined in a management system have been suggested (Frame and Boyd , 1987b) and initial investigations have been made (Younie et al., 1986). 4.5 A B B O T S F O R D F E R T I L I Z E R T R I A L 4.5.1 A n n u a l Yie lds : Annual yields at Abbotsford were considerably lower than those at Oyster River and Agassiz during 1985. Applications of 300 kg N/ha (regimes II, III and IV) significantly increased annual D . M . yield when compared to the 0 N treatment of regime 1 (Table 21 and Figure 30). There were no significant (P > 0.05) differences in annual yield produced by the three regimes receiving 300 kg N/ha/year. Annual yield under the 0 N regime was 67% of the average annual yield produced by regimes II, III and IV, a result very similar to Oyster River where the 0 N produced a relative yield of 69%. Since the swards contained virtually no clover it can be assumed that there was little clover N contribution towards total D . M . yield. Dry matter response to applied N was somewhat low and considerably less than that measured at Oyster River. Response for regimes II, III and IV was 6.0, 5.3 and 6.9 kg D . M . / kg N applied respectively. Such low yields and low D . M . response can be largely attributed to the well draining nature of the soil combined with the exceptionally dry summer. 129 4.5.2 Seasonal Dis tr ibut ion of Y i e l d : Dry Matter yields at each harvest are shown in Table 21 and Figure 31. Growth curves developed by the same method as those for Agassiz are shown in Figure 32. 4.5.2.1 Spring Production: Yie lds at the first harvest on M a y 22 showed no significant difference and growth rates were very similar for all four treatments. Application of N on M a y 22 resulted in twice the yield on June 17 for regimes II, III and IV compared to I which received no N . However, D . M . yields for June 17 were virtually identical for II, III and IV even though II received 150 kg N and III and IV received 100 kg N / ha. Yie ld responses to the May 22 N application at the next harvest (June 17) for II, III and IV were 5.7, 8.2 and 8.5 kg D . M . / kg applied N respectively. The relatively low response of II was due to the short regrowth interval (26 days) which did not allow herbage to respond to higher N levels (150 vs 100 kg / ha). This same type of delayed response was also observed at Oyster River. Growth rates were most rapid throughout Apr i l with little difference between the four treatments. However, N application on May 22 increased growth rates of II, III and IV to a peak in early June of about 67 kg D.M. /ha /day . Growth under regime I peaked in early May at 57 kg D.M./ha/day and declined for the rest of the season. 4.5.2.2 Summer Production: Growth rates declined steadily from spring / early summer peaks to a low in early August of about 2 to 4 kg D.M./ha/day. Growth actually stopped in the late July / early August period and the very low yields 130 harvested on August 22 had accumulated from mid-July. Since this yield was averaged over the regrowth period, zero growth rates are not evident on the growth curves. Yields during the summer were low (93 to 883 kg D.M. /ha) and generally showed poor response to applied N (ranging from 1.2 to 6.1 kg D . M . / kg N applied at previous harvest). With the well draining soil, lack of rainfall and no irrigation, it is not surprising that production was low and N response poor. The effects were similar to Oyster River but more extreme. Irrigation did allow good response at Oyster River when applied. It is also evident that residual effects from previously applied higher N levels were quite small, again due to the lack of moisture. 4.5.2.3 Fall Production: Herbage growth responded to September rainfall to give higher yields on October 5 (510 to 596 kg D.M./ha) for II, III and IV. Growth rates of I did not increase and remained low (2.1 kg D.M./ha/day) in comparison to II, III and IV which responded to previous N applications to give higher growth rates (11.6 to 14.0 kg D.M./ha/day) . August 22 N applications to III and IV showed no response on October 5. Subsequent observations revealed growth into late October, without any visual differences between II, III and IV and with no measurable production. Since the plots were destroyed that winter, no determination of the effect from fall N application could be made for the spring of 1986. The distribution of annual yield at each harvest is shown in Table 22. Fertil ization, in addition to producing higher yields, shifted production from the spring (mean of 74.0%) to the summer (15.7%) and fall (10.4%) when compared to the 0 N regime. However, the distribution of yield for TABLE 21 : YIELD FOR ABBOTSFORD FERTILIZER TRIAL. 1985 FERTILIZER HARVEST DATE TOTAL TREATMENT MAY 22 JUNE 17 JULY 12 AUG 22 OCT 5 YIELD Yield (ka D.M./ha) 1985 I 2414 a 887 b 271c 93 b 95 b 3758 b II 2365 a 1739 a 676 b 165 ab 618 a 5562 a III 2332 a 1707 a 559 b 155 ab 596 a 5350 a IV 2512 a 1732 a 883 a 206 a 510 a 5842 a MEAN YIELD 2406 1516 597 155 455 5128 S.E. OF MEAN 171 71 81 35 120 233 L.S.D. 386 160 183 79 272 528 FIGURE 3 0 : A N N U A L YIELDS FOR ABBOTSFORD FERTILIZER TRIAL A N N U A L YIELD (kg D.M. / ha) I II III IV FERTILIZATION REGIME t—> to F I G U R E 31: Y I E L D AT E A C H H A R V E S T A B B O T S F O R D F E R T I L I Z E R T R I A L , 1 9 8 5 3000 YIELD (kg D.M. / ha) 2 5 0 0 -MAY 22 JUNE 17 JULY 12 AUG 22 OCT 5 H A R V E S T DATE FERTILIZATION REGIME i H II II IV 134 TABLE 22: PERCENT DISTRIBUTION OF ANNUAL YIELD. ABBOTSFORD. 1985 Fertilizer Spring Summer Fall Regime (to June 17) (to Aug 22) (to Oct 5) % of annual yield I 87.8 9.7 2.5 II 73.8 15.1 11.1 III 75.5 13.3 11.2 IV 72.6 18.6 8.8 135 the three N treatments showed litt le difference over the three seasons due to so i l moisture deficits not a l lowing D . M . responses to N to occur and greatly reducing residual effects. For this particular year and site, split applications of N during late summer produced li t t le benefit in terms of total y i e l d and distr ibut ion of y i e ld . In consideration of the extreme dryness of the area, only spring / early summer and possibly fa l l N applications should be considered. N application in early summer may show benefit in terms of mid-summer y ie ld i f enough rainfall occurs at this time to al low for a response before a dry period occurs. 4.6. F O R A G E Q U A L I T Y A l l herbage samples from Agass iz harvested during 1985 and 1986 and samples from Oyster R i v e r during 1985 were analyzed for forage qual i ty. Since replicates were composited it is not possible to conduct statistical analysis. A l l samples were analyzed for crude protein (CP) , acid detergent fibre ( A D F ) , ash, ca l c ium, phosphorus, potassium, magnesium, sodium, copper and manganese. Each of the above are discussed in the f o l l o w i n g pages wi th respect to average annual contents and seasonal distribution of content. Individual harvest data for Agass iz are reported in Append ix 2 whi le Oyster R ive r values and average Agass i z values are presented in the fo l lowing pages. Information on the quality of some local forages is presented in Table 23. 4.6.1. C r u d e Protein Crude Protein (CP) was determined by the block digestion method for total N using a conversion factor of 6.25. The main factor influencing the herbage protein content is the supply of available N in the soil which is TABLE 23 : COMPOSITION OF LOCAL FORAGES Herbage Forage Sources and Locations Component Type Shelford and B.C.M.A. & F. B.C.M.A. & F. B.C.M.A. & F. Peterson (1978) (1983) (1985) (1986) Fraser Valley South Coastal Vancouver Isl. South Coastal Protein {%) Pasture Grass Silage Grass Hay 17.86 (10.6-24.3) 12.9(9.5-15.5) 14.1 (9.4-17.7) 12.87 (4.10-23.90) 11.78 (2.00-26.40) 16.9 (8.6-24) 18.2 (8.3-25.5) A.D.F. (%) Pasture Grass Silage Grass Hay 38.33 (23.4-48.1) 35.8 (19-50) 32.9 (27.4-46.2) 31.7 (26.6-36.4) Calcium (%) Pasture Grass Silage Grass Hay 0.517 (0.30-0.74) 0.48 (0.28-0.69) 0.56 (0.2701.12) 0.54 (0.29-1.15) 0.56 (0.23-1.09) 0.49 (0.2-1.66) 0.42 (0.18-1.83) 0.48 (0.23-0.93) 0.50(0.23-1.50) Phosphorus (%) Pasture Grass Silage Grass Hay 0.347 (0.24-0.47) 0.25 (0.15-0.35) 0.19 (0.13-0.22 0.33 (0.15-0.43) 0.38 (0.24-0.60) 0.32 ( 0.09-1.13) 0.30 (0.18-0.53) 0.33 (0.19-0.53) 0.26 (0.10-0.52) Potassium (%) Pasture Grass Silage Grass Hay 2.68 (1.11-4.69) 2.98 (0.97-4.46) 2.36 (0.79-4.20) 1.98 (0.40-9.38) 2.75 (1.15-3.73) 2.62 (1.07-4.17) TABLE 23: COMPOSITION OF LOCAL FORAGES (continued) Herbage Forage Sources and Locations Component Type Shelford and B.C.M.A. & F. B.C.M.A. & F. B.C.M.A. & F. Peterson (1978) (1983) (1985) (1986) Fraser Valley South Coastal Vancouver Isl. South Coastal Sodium (%) Pasture . . . . Grass Silage - 0.120(0.01-0.676) - 0.119(0.015-0.490) GrassHay - 0.113(0.004-1.19) - 0.101(0.002-0.286) Magnesium (%) Pasture 0.245(0.14-0.42) Grass Silage 0.22(0.20-0.30) 0.24(0.14-0.37) 0.21(0.12-0.36) 0.24(0.13-0.38) GrassHay 0.24(0.18-0.28) 0.28(0.11-0.56) 0.21(0.09-1.10) • 0.25(0.11-0.39) Copper (mg/kg) Pasture 1 5 ( 4 - 5 4 ) Grass Silage 2 1 ( 6 - 3 6 ) 10 ( 0 - 18 ) 8 ( 3 - 39 ) 12 ( 7 - 21 ) GrassHay 14 (4 - 27) 1 0 ( 5 - 24) 6 ( 2 - 13 ) 8 ( 4 - 13 ) Manganese Pasture 148(71 - 302) - 167 (33 - 460) (mg/kg) Grass Silage 156(107 - 200) 121(39 - 398) 159(22 - 774) 97(43 - 175) GrassHay 107(56 -168) 120(30 - 367) 96 (16 - 271) 138 often controlled by fertilizer application rates. Herbage N contents are usually within the range of 1.5 to 4.5% (9.38 to 28.13% Protein) (Whitehead, 1966). As grasses become more mature there is a decline in protein content. Application of N wil l increase grass protein content throughout the life of the plant. Sucli increases do not occur uniformly, but are greatest for young vegetative grass and decline rapidly as the plant reaches hay stages. During periods of rapid growth, application of N may result in fairly large yield increases, leading to lower protein content due to a dilution effect and clue to more mature growth with lower protein content. 4.6.1.1. Averages and ranges: Average protein contents for herbage harvested at Agassiz are given in the Table 24 with values at each harvest reported in Table 52 in Appendix II. Seasonal and average protein contents for Oyster River are presented in Table 26. Values at Agassiz ranged from 11.34 to 27.77% (mean, 22.57%) and for Oyster River from 13.26 to 29.86% (mean, 19.26%). These values are higher than those measured for local pastures (Table 23) which could be attributable to soil N status and growth stage of pasture samples. Fairey (1985a) measured average protein contents under a 4 cut management system at the same site. Orchardgrass averaged 13.5% and perennial ryegrass 13.81% protein. Again, values are lower due to a more advanced growth stage at harvest. Under an 8 cut system values were more similar with several perennial ryegrass cultivars averaging 19.38% protein (Fairey, 1985b). Davis (1961) measured protein content for 40 orchardgrass varieties at the same site and obtained values from about 12 to 21% protein (mean, 15.64%), but at lower levels of N fertilization (97 kg 139 N/ha/year). Values obtained at Oyster River and especially Agassiz are on average higher than local values, partly due to fertilizer application and cutting management. High protein contents are not unknown. Beever et al. (1978) obtained perennial ryegrass herbage with 5.1% N (31.88% protein) in September from growth receiving 80kg N/ha six weeks earlier and Puffe et al. (1984) harvested first cut Italian ryegrass with 27.85% protein content. Johns (1955) obtained levels of 4.54% N (28.38% CP) for vegetative ryegrass. However, it must be noted that increased percent protein is not necessarily an in indication of improved feeding value, since the value may indicate varying proportions of true protein and other nitrogenous Table 24: Average Protein Content of Herbage at Agassiz. 1984 &1985 Cul t ivar Cutting Year Weighted T r e a t m e n t 1984 19 85 Mean * Percent Protein Pra ir ia l L I 22.49 22.42 22.46 Pra ir ia l H F 23.46 23.06 23.26 Sumas L I 22.27 20.84 21.62 S u m a s HF 23.47 23.01 23.24 Barlano L I 21.15 22.32 21.68 Barlano H F 21.51 22.82 22.17 Nor lea L I 20.86 22.30 21.50 Nor lea H F 22.83 24.43 23.63 Weigh ted Mean * 22.37 22.78 22.57 *Nolc: Weighted according to number of samples analyzed. 140 compounds. Nitrate content must be considered as a possible constituent in Agassiz and Oyster River herbage especially in those with higher percent protein. Varietal differences are also important. Griffiths (1958) showed that there are varietal differences in nitrate content of ryegrasses. 4.6.1.2. Agassiz: Average protein content showed little difference between 1984 and 1985 (22.37 and 22.78% respectively). However, there was a greater, difference between cutting treatments with high-frequently cut herbage having a higher protein content (23.09%) than low-infrequently cut herbage (21.83%). This can be attributed both to more mature herbage under L I cutting and a dilution effect from low cut herbage. First cut yields contained more protein under H F cutting (19.42% in 1984 and 25.90% in 1985) than under LI cutting (18.33% in 1984 and 25.58% in 1985). Higher protein content of leaves would be diluted to some extent by lower protein content of stubble under low cutting. Johns (1955) noted an increase in % N as samples were taken at higher levels in a 30cm ryegrass sward. Only slight cultivar differences were evident with Barlano (21.95%) having slightly less protein content than Prairial , Sumas and Norlea (22.93, 22.58 and 22.80% respectively). Fleming and Coulter (1963) noted that perennial ryegrass and orchardgrass had similar N contents at a given stage of growth. Seasonal distribution of protein content generally showed an initial increase during late May / early June followed by decreases during the summer. First cut harvests were lower in protein content in 1984 (18.88%) than in 1985 (25.74%) due to lower dry matter yields in 1985 compared to 1984 (1244 and 1605 kg D . M . / h a respectively). Protein contents were generally lowest during the summer dry periods where 141 moisture was limiting, therefore affecting transfer of N into the root zone (Whitehead, 1970). Annual protein yields are summarized in Table 25 and yields at each harvest arc shown in Table 53 of Appendix II. Average protein yield was higher in 1984 than 1985 due to the significantly higher D . M . yield in that year. Differences were less in terms of protein yield between H F and LI cutting than they were for D . M . yield. Prairial yielded more protein under LI cutting (2300 kg D.M. /ha) than under H F cutting (2131 kg D.M./ha) . Both Sumas and Barlano produced very similar protein yield under LI and Table 25: Average Protein Yield of Herbage at Agassiz. 1984 &1985 Cul t ivar Cutting Year W e i g h t e d T r e a t m e n t 19 84 1985 Mean * K g Protein/ha Pra i r ia l L I 2 8 4 5 1647 2 3 0 0 Pra ir ia l H F 2665 15 96 2131 S u m a s L I 25 1 6 1338 1981 Sumas H F 243 9 1457 1948 Bar lano L I 1986 1374 1708 Bar lano HF 19 6 9 13 74 1672 Nor l ea L I 1474 965 1248 Nor l ea H F 1713 947 1330 Weighted Mean * 2228 1354 1808 *Note: Weighted according to number of samples analyzed. 142 H F cutting (Table 25) while Norlea produced more protein under H F cutting (1330 kg D.M. /ha) than under LI cutting (1248 kg D .M. /ha ) . Cultivar yields were of the same ranking as D . M . yields with the order as follows: Prairial (2200) > Sumas (1961) > Barlano (1688) > Norlea (1298 kg D.M. /ha) . The seasonal pattern of protein production followed the pattern of D . M . production. 4.6.1.3. Oyster River: In general protein content at Oyster River was lower than at Agassiz. The average protein content declined from a high in M a y to lower values in July and August before increasing towards the fall (Table 26). This pattern varied somewhat for the different fertilizer treatments. The high clover plots (regime I) had generally lower protein content throughout the season except at the second cut on June 10. Second cut yields were significantly higher for regimes II, III and IV leading to more mature growth and dilution of the protein content. Regime I produced a lower yield with a higher percent clover and subsequently had a greater protein content. Herbage protein content also showed responses to previous N applications. Regimes II and III had higher contents on August 12 in response to previous N applications and regime IV had higher protein content on September 11 (20.80%) and October 29 (22.81%) in response to the August 12 application of 100 kg N/ha. Protein content on May 2, 1986 was slightly higher for regime I than the other three regimes, possibly in response to the higher clover content (Table 26). There was little difference in clover content between the other three regimes on May 2 even though regime IV produced a significantly greater D . M . yield. Annual and seasonal distributions of protein yield are shown in Table 27. Seasonal TABLE 26: PROTEIN CONTENT FOR OYSTER RIVER FERTILIZATION TRIAL 1985 FERTILIZATION HARVEST DATE YEARLY REGIME MAY 9 JUNE 10 JULY 8 AUG 12 SEPT 11 OCT 29 MEAN May 2 1986 PERCENT PROTEIN 1985 I 21.70 29.86 16.41 17.66 17.28 18.03 20.16 17.52 II 22.34 13.23 18.91 19.08 18.33 18.55 18.41 17.24 III 20.87 19.42 17.30 19.42 17.30 17.84 18.69 16.56 IV 21.84 17.53 17.67 18.06 20.80 22.81 19.79 16.89 :ST MEAN 21.69 20.01 17.57 18.56 18.43 19.31 17.05 TABLE 27: TOTAL PROTEIN YIELD FOR OYSTER RIVER FERTILIZER TRIAL. 1985 FERTILIZATION HARVEST DATE TOTAL REGIME MAY 9 JUNE 10 JULY 8 AUG 12 SEPT 11 OCT 29 YIELD May 2 1986 PROTEIN YIELD (kg P.M. / ha) 1985 I 556 987 82 157 35 45 1862 235 II 615 537 171 323 65 90 1801 319 III 533 757 125 400 82 97 1993 293 IV 599 728 114 222 67 252 1981 384 HARVEST MEAN 575 752 123 275 62 121 308 145 distribution of yield generally followed the distribution of D . M . yield with the exception of the June 10 harvest where the high protein content of herbage under regime I produced more total protein. Annual protein yields were very similar, ranking as follows: III (1993) > IV (1981) > I (1862) > II (1801 kg D . M . / h a ) . Even though regime I produced significantly less D . M . yield than the other three regimes, annual protein yields were very similar due to the relatively high protein yield under regime I on June 10. 4.6.2. A c i d Detergent F i b r e ( A . D . F . ) Average A D F values for herbage cut at Agassiz are presented in Table 28 with values for each harvest presented in Table 54 of Appendix II. Seasonal and average values for Oyster River are presented in Table 29. A D F content ranged from 17.39 to 40.21% (mean, 29.06%) for Agassiz herbage and from 25.10 to 36.01% (mean, 30.81%) at Oyster River. These values are somewhat lower than those reported for local grass hay and grass silage (Table 23) due to the harvesting of more vegetative material. The A D F analytical fraction of the plant contains cellulose, lignin and some of the feed ash (Soest, 1965). With increasing maturity, A D F increases clue to an increasing stem/leaf ratio and an accompanying increase in cellulose and lignin. The low A D F values obtained at Agassiz and Oyster River were associated with lush, vegetative growth. Ahloowalia et al. (1981) obtained Modified A D F (Clancy and Wilson, 1966) values ranging from 26.9 to 30.6% for several tetraploid ryegrasses. 146 Table 28: Average A P F Content of Herbage at Agassiz. 1984 &1985 Cul t ivar Cutting Year W e i g h t e d T r e a t m e n t 1984 19 85 Mean * Percent A D F P r a i r i a l L I 32.35 31.56 31 .99 P r a i r i a l HF 30.76 30.11 30.44 Sumas L I 31 .04 32.63 31.76 Sumas H F 29.88 28.89 29.36 Bar lano L I 28.71 27.44 28.13 Bar lano H F 29.06 26 .59 27.83 Nor l ea L I 27 .39 27.03 27.23 Nor lea H F 27.50 24.17 25.84 Weighted Mean * 29.63 28.44 29.06 *Note: Weighted according to number of samples analyzed. 4.6.2.1. Agassiz: A D F content was lower in 1985 (28.44%) than in 1984 (29.63%), probably due to the lower D M yields obtained in 1985 which were on average less mature than herbage in the previous year. Differences were also evident between LI cutting (29.90%) and H F cutting (28.47% A D F ) which could be attributable to both a higher stubble content and more mature growth under LI cutting. The difference was more apparent for orchardgrass (Table 28). Barlano showed little difference in A D F content TABLE 29 ADF CONTENT FOR OYSTER RIVER FERTILIZATION TRIAL. 1985 FERTILIZATION HARVEST DATE YEARLY REGIME MAY 9 JUNE 10 JULY 8 AUG 12 SEPT 11 OCT 29 MEAN MAY 2 1986 PERCENT ADF 1985 I 25.34 36.01 29.13 34.36 31.02 28.40 30.71 27.93 II 25.48 34.71 29.06 34.81 31.18 27.92 30.52 28.38 III 25.47 35.73 32.89 35.02 33.56 28.19 31.81 27.65 IV 26.44 35.51 27.73 34.64 31.72 25.16 30.20 28.72 HARVEST MEAN 25.68 35.49 29.70 34.71 31.87 27.42 28.17 148 between the two cutting treatments. Nor lea only exhibited a difference in 1985 where H F cutt ing produced more vegetative growth. L I cutting resulted in zero growth at times for Norlea , leading to more dead material with a higher A D F content. C u l t i v a r d i f ferences were also apparent w i t h the perennia l ryegrasses (27.20%) having lower A D F contents (thus higher digest ibi l i ty) than the orchardgrasses (30 .71%) . F a i r e y (1985a) noted s i m i l a r differences wi th orchardgrass producing consis tent ly lower dry matter digest ib i l i t ies than perennial ryegrass. The seasonal distr ibution of A D F content showed the same general trend in both years being affected by D M yie ld and cl imate. First cut harvests were general ly vegetative and therefore had fa i r ly low A D F values. Second cut harvests y ie lded more D M wi t l i higher ratios of r ep roduc t ive to vegetat ive t i l le rs resu l t ing in higher percent A D F . Subsequent cuts were more vegetative and therefore produced more digest ible herbage. Peak A D F content was reached during mid-summer even though most growth was vegetative. Summer growth contained herbage wi th a higher proportion of dead material which would lead to higher A D F values and lower digestibil i ty (Shcehan et al . , 1985). When fall regrowth occurred in response to rainfall , the A D F content declined from the mid-summer peak. However , fall A D F % was generally higher than the early spring harvests. Swift et a l . (1952) noted a seasonal decl ine in the digest ibi l i ty of orchardgrass from spring to fall measured from successive cuttings, but made no mention of increased digest ibi l i ty for autumn regrowth. 14 4.6.2.2. Ovsler River Differences in A D F content between fertilization treatments were slight. Application of N tended to lower A D F values. A D F content was slightly lower for regimes II, III and IV compared to I as a result of the May 9 application. Regime IV also showed a lower A D F content on October 29 following the August 12 fertilization. Average percent A D F over the season showed little difference between treatments. Regime III had a slightly higher A D F content (31.81%) compared to the other three regimes (mean, 30.48%) due to higher values from June to October. Seasonal trends in A D F content were similar to those at Agassiz with lower values at first harvest. Second harvest A D F content was higher due to more mature growth. A D F content then decreased for July since growth was more vegetative and increased during the dry summer months of August and September. Vegetative fall regrowth then showed a lower A D F content. 4 . 6 . 3 . Ash In any experiment that reepjires herbage mass to be determined (whether by direct or indirect methods), there is some degree of sample contamination. Johnson (1978) noted that samples can be contaminated by soil , animal excreta or fertilizer which can result in analytical errors through enrichment with minerals or weight errors. Fecal contamination did not occur on any of the Agassiz, Abbotsford or Oyster River plot trials. Fertilizer was always applied after sampling and thus was not a factor. Disc meter samples were taken from areas not contaminated by feces and always at least two weeks after any fertilizer applications. Soil is probably the only source of contamination present in the Agassiz and Oyster River 150 samples. Mole hills were a problem at both sites resulting in occasional resampling if one of these objects was encountered with the plot harvester. Dust was a definite problem at Agassiz since the plots were located 15m from a gravel road which resulted in the plots being frequently covered. As can be expected, the problem was much more severe during the dry months. No visible dust contamination occurred at Oyster River. Average ash contents at Agassiz and Oyster River were very similar (8.92 and 9.08% respectively). 4.6.3.1 Averages and ranges: Average ash values for herbage harvested at Agassiz is given in Table 30. Values at each harvest are reported in Appendix II, table 55. Seasonal and average ash values for Oyster River are presented in Table 31. Values at Agassiz ranged from 6.76 to 9.99% (mean, 8.92%) and values at Oyster River ranged from 7.92 to 9.85% (mean, 9.08%). Plot samples that were known to be contaminated (i.e. by mole hills) were discarded, but a few subsamples were collected and analyzed for ash, giving values from 11.23 to 23.43% Ash. These contaminated samples had a brownish tinge after combustion that was not evident in clean samples. The colour difference was even noticeable at values of 11 to 12% ash. Herbage values with ash contents in excess of 10% are generally considered to be contaminated with soil, although this value will vary with herbage species and growth stage. Van Ripen and Smith (1959) noted a decrease in ash content with advancing maturity of bromegrass with values ranging from 10.15 to 12.16% ash. Lush (1938) measured an average ash content of 10.30% for pasture during an eight year investigation. No mention of soil contamination was made in either of 151 these articles. More recent studies have been made by Nicol l (1980 and 1982) investigating ash content in relation to soil contamination. Samples of ryegrass with 10.2% ash were fell, to be contaminated with soil with unconlaminatcd samples having an average ash content of 7.51%. Machine washing was felt to remove surface soil contamination without affecting estimates of in vitro D M digestibility or organic matter digestibility. Table 30: Average Ash Content of Herbage at Agassiz. 1984 &1985 Cul t ivar Cutting T r e a t m e n t Year W e i g h t e d 19 84 1 9 85 Mean * Percent Ash Pra ir ia l L I 9. .19 8. .19 8. .74 Pra i r ia l H F 8. .81 8. .34 8. .58 S u m a s L I 9. .34 8. .53 8, .97 S u m a s H F 8. .92 8. .26 8, .59 Bar lano L I 9. .19 9, .18 9 .19 Bar lano H F 9. .16 9, .30 9 .23 Nor l ea L I 9, .38 8, .81 9 .13 Nor lea H F 9. .25 9 .06 9 .16 Weighted Mean * 9. .13 8. .70 8 .92 *Nole: Weighted according to number of samples analyzed. However, machine washing did lead to loss of herbage ash which may affect analysis of major and minor mineral elements. In light of the above values, the Agassiz and Oyster River herbage may be contaminated with soil but the exact degree is unknown since values can vary between T A B L E 31: ASH CONTENT FOR OYSTER RIVER FERTILIZATION TRIAL. 1985 FERTILIZATION HARVEST DATE YEARLY REGIME MAY 9 JUNE 10 JULY 8 AUG 12 SEPT 11 OCT 29 MEAN MAY2 1986 PERCENT ASH 1985 I 8.84 9.03 9.66 9.81 8.91 9.01 9.21 8.05 II 9.38 9.55 8.68 9.85 8.20 9.27 9.15 8.29 III 8.70 7.97 8.59 9.64 8.32 9.16 8.73 8.25 IV 9.55 9.21 9.33 9.62 7.92 9.75 9.23 8.43 HARVEST MEAN 9.12 8.94 9.07 9.73 8.34 9.30 8.26 153 species and growth stages. Machine washing of samples prior to analysis was not considered due to mineral losses from this process. Ash values at Agassiz and Oyster River did not vary greatly over the seasons and it is felt that any soil contamination was not that great. 4.6.4. C a l c i u m : 4.6.4.1 Averages and ranges: Average calcium (Ca) content for herbage harvested at Agassiz is given in Table 32. Values at each harvest are reported in Appendix II, Table 56. Seasonal and average Ca contents for Oyster River are presented in Table 33. Values at Agassiz ranged from 0.268 to 1.264% C a (mean, 0.572%) while values at Oyster River ranged from 0.585 to 1.400% (mean, 0.840). Values for Agassiz are similar to those from local pasture, hay and silage analysis (Table 23) but those for Oyster River are somewhat high partly as a result of the high clover content of the regime I plots. Ca levels in grass are usually in the range of 0.4 to 1.0% and with legumes, contents in excess of 3% are not unusual (Whitehead, 1966). Content of New Brunswick forages is generally lower with averages for several sites ranging from 0.30 to 0.35% Ca (Nicholson, 1982). Van Riper and Smith (1959) measured the Ca content of bromegrass in the vegetative stage and found ranges from 0.39 to 0.61%. C a content then decreased with advancing maturity from the boot to dough stage. 4.6.4.2. Agassiz: Average Ca content showed little difference between 1984 and 1985 (0.564 and 0.593% respectively). No discernible differences in C a content due to cutting treatment were evident (Table 32) except for Prairial which showed a slightly higher Ca content under LI cutting (0.527%) compared to H F cutting (0.489%). Differences between species were apparent with 154 perennial ryegrass (0.644) having a higher C a content than orchardgrass (0.519%), a result which has been observed in other studies (Thomas et al., 1952 and Puffe et al., 1984). The seasonal distribution of calcium content Table 32: Average Calcium Content of Herbage at Agassiz. 1984 & 1985 Cul t ivar Cutting Y e a r Weighted T r e a t m e n t 19 84 1985 Mean * Percent Calcium Pra i r ia l L I 0.519 0.537 0.527 Pra i r ia l H F 0.472 0.505 0.489 S u m a s L I 0.517 0.550 0.532 Sumas H F 0.506 0.563 0.535 Barlano L I 0.619 0.691 0.652 Barlano H F 0.616 0.695 0.656 Nor lea L I 0.626 0.643 0.634 Nor l ea H F 0.669 0.596 0.633 Weighted Mean * 0.564 0.593 0.572 *Note: Weighted according to number of samples analyzed. generally displayed an increase in Ca content from May to September with a slight decline in October. 4.6.4.3. Oyster River: Ca content was generally higher than at Agassiz. Values showed a steady increase from May to October (Table 33). The annual averages for TABLE 33: CALCIUM CONTENT FOR OYSTER RIVER FERTILIZATION TRIAL. 1985 FERTILIZATION HARVEST DATE YEARLY REGIME MAY 9 JUNE 10 JULY 8 AUG 12 SEPT 11 OCT 29 MEAN MAY 2 1986 1985 I 0.585 0.665 II 0.600 0.590 III 0.625 0.665 IV 0.645 0.600 HARVEST MEAN 0.614 0.630 PERCENT CALCIUM 1.215 1.365 1.350 0.805 0.820 . 0.755 0.910 0.790 0.750 0.805 0.850 0.845 0.934 0.956 0.925 1.400 1.097 0.625 0.865 0.739 0.465 0.945 0.781 0.450 0.715 0.743 0.435 0.981 0.494 156 regimes II, III and IV were very similar (0.739, 0.781 and 0.743% Ca respectively). C a content for regime I (1.097%) was higher than the other three regimes (mean, 0.754%) due to the high clover content. Clover has a higher C a content than grass and increases the Ca content of mixed swards in proportion to the amount of clover. The first cut in May 1986 exhibited a similar trend with regime I herbage having a higher C a content (0.625%) than the other three regimes (mean, 0.450%). 4.6.5 Phosphorus: 4.6.5.1 Averages and ranges: Under normal agricultural conditions the phosphorus (P) content of grasses and clover varies less widely than most other minerals. Values outside of the range of 0.2 to 0.5% P are unusual (Whitehead, 1966). Average P content for herbage harvested at Agassiz is given in Table 34 with values at each harvest reported in Table 57 of Appendix II. Seasonal and average P contents for Oyster River are presented in Table 35. Values at Agassiz ranged from 0.219 to 0.491% P (mean, 0.357%) and for Oyster River ranged from 0.237 to 0.374% (mean, 0.289). These values are within the normal range given by Whitehead (1966). Averages and ranges at both Agassiz and Oyster River are close to local average P values and are well within the ranges given (Table 23). P content declined with advancing maturity which is evident from local forage analyses having lower averages for grass hay and grass silage than pasture (Van Riper and Smith, 1959; Fleming and Coulter, 1963). Herbage analysis of New Brunswick pastures revealed similar P contents ranging from 0.25 to 0.30% at several sites (Nicholson, 1982). Ryegrass cultivars showed ranges 157 from 0.31 to 0.40% P in Ireland (Culleton and Fleming, 1983) and Puffe et al. (1984) reported a range from 0.22 to 0.45% P in several grasses in Germany . Table 34: Average Phosphorus Content of Herbage at Agassiz. 1984 &1985 Cul t ivar Cutting Year ^Weighted T r e a t m e n t 1 9 84 1 9 85 Mean * Percent Phosphorus Pra ir ia l L I 0.375 0.351 0.364 Pra ir ia l H F 0.348 0.352 0.350 Sumas L I 0.356 0.330 0.344 Sumas H F 0.325 0.333 0.329 Barlano L I 0.363 0.377 0.369 Barlano H F 0.377 0.367 0.372 Nor lea L I 0 .362 0.371 0.366 Nor lea H F 0.395 0.347 0.371 Weighted Mean * 0.361 0.352 0.357 *Nole: Weighted according to number of samples analyzed. 4.6.5.2 Agassiz: There appeared to be no consistent differences in P content due to cutting treatments and the averages for 1984 and 1985 were quite similar (Table 34). Differences were evident between ryegrasses (0.370% P) and orchardgrasses (0.347% P) with the difference fairly consistent throughout the seasons. No references were found which indicate differences in P TABLE 35: PHOSPHORUS CONTENT FOR OYSTER RIVER FERTILIZATION TRIAL.1985 FERTILIZATION REGIME 1985 I II III IV MAY 9 JUNE 10 HARVEST DATE YEARLY JULY 8 AUG 12 SEPT 11 OCT 29 MEAN 0.298 0.327 0.297 0.321 HARVEST MEAN 0.311 PERCENT PHOSPHORUS 0.298 0.294 0.278 0.320 0.307 0.237 0.243 0.262 0.295 0.270 0.283 0.297 0.291 0.258 0.249 0.267 0.298 0.262 0.286 0.266 0.276 0.285 0.296 0.374 0.308 0.294 0.279 0.274 0.307 MAY 2 1986 0.304 0.295 0.301 0.337 0.309 u i oc 159 content between orchardgrass and perennial ryegrass. Culleton and F leming (1983) noted differences between perennial ryegrasses and Italian ryegrasses and Plummer (1953) had indications of differences in P content between orchardgrass, reed canary grass and timothy. There were no consistent seasonal trends in P content at Agassiz in either 1984 or 1985. 4.6.5.3 Ovster River: Average P content at each harvest showed consistent trends over the growing season and no consistent responses to N application were evident. P content was higher under regime IV on October 29, 1985 and May 2, 1986 (0.307 and 0.337% P respectively) than under the other three regimes which had average P contents of 0.286% on October 29 and 0.300% on May 2. This could have been due to a residual effect of N application on August 12, 1986. However, N application has been shown to decrease P content (MacLeod et al., 1964) possibly as a result of N increasing yield resulting in P dilution (Lightner et al., 1981) The high clover content of regime I did not affect the average P content of herbage on these plots, although they were higher than the other three regimes on July 8 and September 11. In general grasses and legumes contain similar P levels (Whitehead, 1966) but Plummer (1953) showed red clover to have higher levels than orchardgrass. 160 4 . 6 . 6 P o t a s s i u m : 4.6.6.1 Averages and ranges: Average potassium (K) content from herbage harvested at Agassiz is given in Table 36. Values at each harvest are presented in Table 58 in Appendix II. The seasonal and average K contents for Oyster River are shown in Table 37. Values at Agassiz ranged from 1.37 to 4.96% (mean, 3.23) and from 1.35 to 4.08% (mean, 2.65%) at Oyster River. Values at Oyster River were within ranges given for local forages with an average value very similar to those for grass hay and grass silage (Table 23). The range of values at Agassiz was also within ranges given for local forages but the average at Agassiz was somewhat higher (about 0.3 to 0.4%) than local averages. Most values at Agassiz and Oyster River were within the usual range of 1 to 4% (Whitehead, 1966). Values reported by Nicholson (1982) for New Brunswick forages were somewhat lower, ranging from 2.2 to 2.65%. Puffe et al. (1984) examined several grasses in Germany and obtained a range from 1.70 to 3.64% K . Culleton and Fleming (1983) examined several ryegrasses, with K levels ranging from 2.92 to 3.24 %. Under grazing situations, K levels can be significantly affected by the grazing animals. Tallowin and Brookman (1988) measured pasture K levels in A p r i l before grazing and found levels to be fairly uniform, averaging 1.9% K . However, after continuous grazing, significantly lower K levels were found in grazed areas (1.8%) compared to a level of 2.4% in rejected areas . The main factor affecting K herbage content is the amount of available K in the soil. This is influenced by a number of factors including the amount of K fertilizer, soil type, p H and C a and M g status (MacLeod et 161 al., 1964; Whitehead, 1966; Oohara et al, 1981; Culleton and Fleming, 1983). Low K levels can reduce herbage production but it is unlikely for herbage to contain insufficient K for animal requirements. For grass growth de Wit et al. (1963) felt that a critical level was 1% K while MacLeod (1965) considered that the optimum K content of the three grass species (timothy, bromegrass and orchardgrass) he examined was above 2%. Values at Oyster River did not drop below 1%, but values lower than 2% were recorded at both sites in 1985. Table 36: Average Potassium Content of Herbage at Agassiz. 1984 &1985 Cul t ivar Cutting Y e a r Weighted T r e a t m e n t 1984 1985 Mean * Percent Potassiu m P r a i r i a l L I 3 .90 2.21 3.13 P r a i r i a l H F 3 .51 2.79 3.15 S u m a s L I 3 .98 2.52 3.32 S u m a s H F 3 .47 2.93 3.20 Barlano L I 3 .10 2.77 2.95 Barlano HF 3 .43 3.12 3.28 Nor lea L I 3 .57 2.17 2.95 Nor lea H F 3 .78 3.63 3.71 Weighted Mean * 3 .59 2.84 3.23 *Notc: Weighted according to number of samples analyzed. 162 4.6.6.2 Agassiz: K content was greater for 1984 (3.59%) than 1985 (2.84%) even though plots received the same amount of K fertilizer in both years. Lime was applied before harvests in 1985 and this could account for the decrease in K content since application of lime has been shown to reduce herbage K content (MacLeod et al., 1964). Cutting treatment had no effect on K content for orchardgrass but H F cutting increased K content for both perennial ryegrass cultivars, possibly due to more immature leafy grass being harvested under H F cutting (Fleming and Coulter, 1963; ap Griffith and Walters, 1966). Slight differences between cultivars were evident with the ranking as follows: Norlea (3.41%) > Sumas (3.25%) > Prairial (3.14%) > Barlano (3.13%). ap Griffith and Walters (1966) noted little differences between grass species but within a species twofold differences were evident between cultivars. Culleton and Fleming (1983) noted little cultivar differences but found differences between ryegrass species. Overall , seasonal K content tended to decline as the season progressed after reaching a peak yield; a result observed by several researchers (ap Griffith and Walters, 1966; Culleton and Fleming, 1983; Puffe et al., 1984). Increases were generally evident in the fall. 4.6.6.3 Oyster River: Average K content was higher for the high N plots (2.80%) compared to the low N plot (2.17%). The same trend was evident for the May 2, 1980 harvest with high N plots containing more K (3.04%) than the low N plots (2.56%). The result can be attributed to N application since addition of N has been found to increase K levels when there is an adequate soil supply of K (Kemp, 1960; Wolton, 1963). The high clover content of the low N TABLE 37: POTASSIUM CONTENT FOR OYSTER RIVER FERTILIZATION TRIAL. 1985 FERTILIZATION HARVEST DATE YEARLY REGIME MAY 9 JUNE 10 JULY 8 AUG 12 SEPT 11 OCT 29 MEAN MAY 2 1986 PERCENT POTASSIUM 1985 I 3.56 1.35 2.26 2.18 1.74 1.92 2.17 2.56 II 3.75 3.65 2.41 3.11 2.23 2.36 2.92 2.92 III 3.41 2.99 1.98 2.64 1.70 2.33 2.51 3.06 IV 3.26 4.08 2.41 2.96 1.79 3.35 2.98 3.13 HARVEST MEAN 3.50 3.02 2.27 2.72 1.87 2.49 2.92 164 plots probably did not increase K content since there is little difference between grasses and clover under conditions of adequate soil K . Over the season K content steadily declined from a May peak (3.5%) to a September low (1.87%) followed by an increase to 2.49% in late October. Seasonal patterns were evident in response to the different fertilization regimes. For example, K levels were higher in response to applied N on August 12 for regime II (3.11%) and on Oct. 29 for regime IV (3.35%). Spring content in 1986 was higher for regime IV (3.13%) than regimes II and III (mean, 2.99%) in response to the August 12, 1985 N application. 4.6.7 Sodium: 4.6.7.1 Averages and ranges: Herbage sodium (Na) levels have received much less attention than other minerals. Herbage requirements are not well defined, but it has been reported that levels below 0.13% can result in N a deficiency. Average N a content for herbage harvested at Agassiz is given in Table 38 with values at each harvest reported in Table 59 in Appendix II. Seasonal and average N a contents for Oyster River are presented in Table 39. Values at Agassiz ranged from 0.053 to 0.319% (mean, 0.145%) and from 0.097 to 0.350% (mean, 0.185%) at Oyster River. A l l values were within the ranges measured for local forages with means at Agassiz and Oyster River being slightly higher (Table 23). N a contents for New Brunswick forages were generally much lower 0.010 to 0.015% (Nicholson, 1982). Values in the literature report N a content varying by more than a 1000-fold from 0.002 to 2.12% and any values between 0.05 and 1% would not be considered unusual (Whitehead, 1966). Puffe et al. (1984) reported ranges from 0.003 to 0.092% in Germany and Culleton and Fleming (1983) 165 obtained values from 0.10 to 0.41% for ryegrasses in Ireland. Jones (1963) reported mean values from 0.01 to 0.03% for several African grassland species. The K content of herbage is largely affected by Na and K content of the soil, with the soil Na content being greater when soil K status is low (Henkens, 1965). Table 38: Average Sodium Content of Herbage at Agassiz. 1984 &1985 Cul t ivar Cutting Year W e i g h t e d T r e a t m e n t 1984 19 85 Mean * Percent Sodium P r a i r i a l L I 0.243 0.199 0.223 P r a i r i a l H F 0.235 0.196 0.216 Sumas L I 0.119 0.082 0.102 Sumas H F 0.133 0.087 0.110 Bar lano L I 0.203 0.184 0.194 Barlano H F 0.158 0.197 0.178 Nor lea L I 0.071 0.059 0.067 Nor lea H F 0.068 0.049 0.059 Weighted Mean * 0.156 0.134 0.145 *Notc: Weighted according to number of samples analyzed. 4.6.7.2 Agassiz: The most significant difference in Na content was variation between cultivars with Prairial (0.219%) and Barlano (0.186%) having much higher levels than Sumas (O.106%) and Norlea (0.063%). The exceptionally low Na TABLE 39: SODIUM CONTENT FOR OYSTER RIVER FERTILIZATION TRIAL. 1985 FERTILIZATION REGIME 1985 I II III IV MAY 9 JUNE 10 HARVEST DATE JULY 8 AUG 12 SEPT 11 OCT 29 0.105 0.119 0.114 0.097 0.154 0.155 0.255 0.136 PERCENT SODIUM 0.137 0.276 0.350 0.205 0.154 0.186 0.310 0.198 0.179 0.173 0.217 0.171 0.138 0.164 0.217 0.217 YEARLY MEAN 0.145 0.179 0.244 0.171 MAY 2 1986 0.193 0.160 0.141 0.163 HARVEST MEAN 0.109 0.175 0.242 0.212 0.185 0.184 0.164 167 content of Norlea was also demonstrated by ap Griffith and Walters (1966). 'Latar' orchardgrass was also found to be low in Na. They noted large and consistent differences between grass genera and classified grasses as having a low or high potential for Na content. A grass with a potentially high N a content will show a wide range in this element depending on the environment. One with a low Na potential will always remain low, not unaffected by environment but affected within a more restricted range. This is especially true for Norlea and less so for Sumas. Culleton and Fleming (1983) showed differences in N a contents between early, mid-season and late perennial ryegrass and Italian ryegrasses with various cultivar differences within a species. Differences in N a content between cutting treatments and years were not very large. L I cutting yielded slightly higher Na contents (except for Sumas) and average Na content in 1985 was only 0.022% below the 1984 average. No consistent seasonal trends were evident. 4.6.7.3 Ovster River: Plots receiving high N (regimes II, III and IV) showed higher average N a contents (mean, 0.198%) than the low N plots (0.145%). This can be attributed to the increase of Na content following N application without K (Stewart and Holmes, 1953; ap Griffith and Wallers, 1966). Application of K causes significant reduction in Na content (Kemp, 1960; Hemingway, 1961). On average the Na content increased to mid-summer and then decreased towards the fall. Regime III maintained a consistently higher N a level from June 10 onwards and then had the lowest content on May 2, 1986 (0.141%). Such differences are not readily explainable since regimes II, III and IV all received the same annual N application and 168 regime II received higher N application than I I I when Na contents of regime III plots showed large increases. Regime I showed a higher Na content (0.193%) on May 2, 1986 than the other three regimes (mean, 0.155%) probably due to a higher Na content in the clover portion. 4.6.8 M a g n e s i u m : 4.6.8.1 Averages and Ranges: The main significance of herbage Magnesium (Mg) content is its relation to Hypomagnesemia of grazing animals. Average M g content for herbage harvested at Agassiz is given in Table 40 with values at each harvest reported in Table 60, Appendix I I . Seasonal and average Mg contents for Oyster River are presented in Table 41. Values at Agassiz ranged from 0.137% to 0.386% (mean, 0.233%) and from 0.136% to 0.367% (mean 0.251%) for Oyster River. These values are comparable to usual herbage M g contents which range from 0.08 to 0.30% (Whitehead, 1966). Mean values at Agassiz are quite close the those obtained from analysis of local forages while the mean value at Oyster River is slightly higher (Table 23). Ranges at Agassiz and Oyster River are well within the ranges given for local forages. M g contents of New Brunswick forages are somewhat lower, ranging from 0.14 to 0.17% which is considered too low for lactating beef and dairy cattle (Nicholson, 1982). Although no direct relationship between the incidence of Hypomagnesemia and herbage M g contents has been demonstrated, several researchers have suggested a value of 0.20% M g as being critical (Kemp, 1960; Wolton, 1963). During fall grazing, higher values (0.25% Mg) have been suggested as being critical (Todd, 1966) and similar values have been suggested throughout the year when K and N levels are high (Metson et al., 1966). 169 M g content of herbage is affected by several factors including: soil type, herbage species, ferti l izer appl icat ion, season and climate. Deficiencies usually occur on three general soil types: (i) light sandy soils with low M g contents; (ii) acid soils and (iii) soils with high K levels (Whitehead, 1966). Spring M g contents are often low, usually in association with colder and wetter weather. Rapid increases in temperature after a cool spell are often associated with rapidly increasing K levels and slower increases in M g and C a resulting in higher K : ( C a and Mg) ratios which are associated with higher incidence of hypomagnesemia (Kemp, 1960; Grunes, 1967). Table 40: Average Magnesium Content of Herbage at Agassiz. 1984 & 1 9 8 5 Cul t ivar Cutting T r e a t m e n t Year 19 84 1985 W e i g h t e d Mean * Percent Magnesium Pra ir ia l L I 0.241 0.221 0.232 Pra ir ia l H F 0.218 0.213 0.216 Sumas L I 0.259 0.252 0.256 S u m a s H F 0.244 0.249 0.247 Bar lano L I 0.256 0.277 0.266 Barlano H F 0.217 0.25 8 0.238 Nor lea L I 0.224 0.217 0.221 Nor lea HF 0.194 0.198 0.196 Weigh ted Mean * 0.231 0.235 0.233 Note: Weighted according to number of samples analyzed. TABLE 41: MAGNESIUM CONTENT FOR OYSTER RIVER FERTILIZATION TRIAL. 1985 FERTILIZATION REGIME 1985 I II III IV MAY 9 0.136 0.143 0.137 0.142 JUNE 10 HARVEST DATE JULY 8 AUG 12 SEPT 11 OCT 29 0.186 0.194 0.215 0.181 PERCENT MAGNESIUM 0.310 0.268 0.296 0.270 0.367 0.304 0.325 0.327 0.317 0.291 0.291 0.270 0.292 0.252 0.273 0.244 YEARLY MEAN 0.268 0.242 0.256 0.239 MAY 2 1986 0.174 0.158 0.165 0.164 HARVEST MEAN 0.140 0.194 0.286 0.331 0.292 0.265 0.165 171 4.6.8.2 Agassiz: M g values in 1984 and 1985 were very similar (0.231 and 0.235% respectively). Differences were evident between cutting treatments with L I cutting (0.245%) producing higher herbage levels than H F cutting (0.225%). Such differences could be attributable to higher M g content in the stubble than leaves which would be harvested by low cutting. Cultivar differences were also evident with Sumas and Barlano being very similar (0.251 and 0.250% respectively) but higher than Prairial (0.223%) which in turn was greater than Norlea (0.206%). Differences in grass species have been noted: Thomas et al. (1952) found Timothy (P hie urn pratense) to be lower in M g than ryegrass and orchardgrass while Todd (1961) found the reverse. Currier et al. (1981) examined four orchardgrass cultivars with the intent of breeding for higher M g content and found significant differences in herbage M g levels. A l l first cut harvests and some second harvests had M g contents below 0.20% typical of the low spring levels. In general M g levels increased as the season progressed, a result in agreement with other work (Hemingway, 1962; Todd, 1961). Samples taken in late October showed a slight increase under H F cutting. 4.6.8.3 Oyster River: First and second harvest M g contents were generally low followed by increasing levels through the season to August 12 followed by a decline on September 11 and late October. Average M g content was higher for 172 regime I than for regimes II, III and IV (mean, 0.246%) clue to the higher clover content. Clover has higher M g levels than grass, and forage with a high clover component has higher M g levels when compared with a pure grass sward (Thomas et al., 1952; Todd, 1961; Hemingway, 1962; Oohara et al, 1981; Nicholson, 1982). M g levels for regime I became higher from July 8 onwards as the clover component increased through the summer. Variations in M g content for regimes II, III and IV were erratic and no discernable patterns were evident other than increasing M g content as the season progressed. In general it has been found that N application can increase M g levels but that large N application in association with heavy K dressings decrease herbage M g content. K application causes decreases in herbage M g (Whitehead, 1966). Herbage M g levels on M a y 2, 1986 showed no difference between regimes II, III and IV (mean, 0.162%). However, M g content of regime I was greater (0.174%) due to the much higher clover component on those plots. 4.6.9 C o p p e r ; 4.6.9.1 Averages and ranges: Average copper (Cu) content from herbage harvested at Agassiz is given in Table 42. Values at each harvest are reported in Table 61 in Appendix II. Seasonal and average Cu contents for Oyster River are presented in Table 43. Values at Agassiz ranged from 8.7 to 24.2mg/kg (mean of 14.6) and from 7.2 to 20.2 (mean of 14.8) for Oyster River. These values are in general agreement with analysis of local hay and pasture samples (Table 23). Values are higher than those reported for grass silage and hay, due to the fact that more mature, stemmy grass is lower in Cu than young leafy material (Adams and Elphick, 1956). Copper content of 173 pastures in other areas arc much lower than those reported at Agassiz and Oyster River. Metson et al. (1979) measured values ranging from 7.3 to 13.4 mg/kg on several sites in New Zealand and Nicholson (1982) showed a range from about 5 to 6 mg/kg at several sites in New Brunswick. Such differences can be attributed to a number of factors. It has been demonstrated that Cu levels can be affected by: soil types, growth stage, botanical composition, fertilizer treatment, soil temperature, climate and advancing season (Adams and Elphick, 1956; Hemingway, 1962; Reay and Marsh, 1976; Reddy et al., 1981). Table 42: Average Copper Content of Herbage at Agassiz. 1984 & 1985 Cul t ivar Cutting Year W e i g h t e d T r e a t m e n t 1984 1985 Mean * Copper Content (mg/kg) P r a i r i a l L I 15.7 15.1 15. .4 Pra i r ia l H F 16.6 16.9 16. .8 S u m a s L I 14.6 14.8 14, .7 Sumas H F 14.9 16.7 15. .8 Bar lano L I 13.1 13.2 13, .1 Bar lano H F 11.5 14.4 13, .0 Nor l ea L I 11.8 14.9 13 .2 Nor lea H F 1 1.9 15.8 13 .9 Weighted Mean * 13.9 15.4 14 .6 y-Note: Weighted according to number of samples analyzed. TABLE 43: COPPER CONTENT FOR OYSTER RIVER FERTILIZATION TRIAL 1985 FERTILIZATION HARVEST DATE YEARLY REGIME MAY 9 JUNE 10 JULY 8 AUG 12 SEPT 11 OCT 29 MEAN COPPER CONTENT (mg/kg) 1985 I 9.2 17.6 15.5 16.7 14.4 13.7 14.5 ll 10.1 17.4 20.2 15.7 15.5 15.8 15.8 III 7.2 17.2 16.8 12.7 14.9 16.4 14.2 IV 10.0 17.6 14.7 12.8 16.1 16.1 14.6 RVEST MEAN 9.1 17.5 16.8 14.5 15.2 15.5 MAY 2 1986 HARVEST 11.1 10.2 9.1 13.0 10.9 175 4.6.9.2 Agassiz: Early harvests in eacli year were generally lower in Cu than later harvests, possibly clue to a dilution factor and lower Cu levels in more mature grass. Orchardgrass (mean, 15.8mg/kg) has a higher Cu content than perennial ryegrass (mean, 13.3). Prairial (mean, 16.1) was higher than Sumas (mean, 15.3) but there were no discernable differences between Norlea and Barlano. High-Frequent cutting (mean, 15.0) produced herbage with a higher Cu content than Low-Infrequent cutting (mean, 14.1). Such cutting treatment effects could be caused by varying distribution of Cu in the herbage. Higher Cu levels in the leaves might be diluted somewhat by lower levels in stubble resulting in lower values under low cutting than high cutting. A more probable explanation is that under L I cutting, herbage was more mature and thus more likely to have lower Cu levels than the less mature herbage under H F cutting. Soil contamination would most l ikely create a reverse situation since contamination is higher at lower cutting heights. Seasonal distribution was somewhat erratic, but there seemed to be an initial increase to a maximum followed by a decrease during July and/or August followed by another increase and subsequent decrease. To attribute this trend to one or a few factors is difficult since several interrelating factors are invo lved . Other research into seasonal distribution has produced variable results. Metson et al. (1979) showed very little seasonal variation in New Zealand whereas Reid et al. (1970) showed a range of 12 to 18mg/kg. Under a four cut silage system Hemingway (1962) noted an increase in C u content as the season progressed. Obviously the seasonal distribution of Cu content is complex 176 and examination of many factors over several years is required to obtain a better understanding. 4.6.9.3 Oyster River: First harvests in each year were lower in Cu than later harvests, possibly due to a dilution effect from higher yields early in the season and lower Cu content of more mature grass. Differences in Cu levels between the low N / high clover plots and the high N / low clover plots were not evident. Yearly averages for all four regimes were fairly close and may be more related to fertilizer regime than clover content. Application of N has been shown to increase Cu content in grass swards (Stewart and Holmes, 1953; Hemingway, 1962) and grass / clover swards (Havre and Dishington, 1962). Regime II showed a higher Cu content on July 8 possibly in response to the 300 kg N/ha that had already been applied. Also on July 8 regime III (250 kg N/ha) had a higher Cu level than regime IV (200 kg N/ha). The first harvest on May 2, 1986 indicated a higher level for regime IV (13.0mg/kg) than the other three regimes (mean, lO. lmg/kg) possibly due to the N application of 100 kg/ha on August 12, which may suggest a higher uptake and storage in the fall as a result of N or a generally larger root mass and/or higher absorption in spring. No consistent seasonal trends were evident. The influence of clover contents, fertilizer regime, climate and growth stage would mask any seasonal trends. 177 4.6.10 Manganese: 4.6.10.1 Averages and ranges: Average manganese (Mi l ) content for herbage harvested at Agass iz is given in Table 44. Values at each harvest are reported in Appendix II, Table 62. Seasonal and average M n contents for Oyster R i v e r are presented in Table 45. Values at Agass iz ranged from 88 to 137 mg/kg (mean, 116) and from 47 to 160 (mean, 93) for Oyster River . These ranges are we l l wi thin the ranges reported for local forage analyses (Table 23). Average values are at the lower end of averages for local forages due to the absence of high M n values found in some local samples (for example a maximum of 774 mg/kg on Vancouver Island and 398 mg/kg in the Fraser Va l l ey ) . In general herbage M n contents are within the range of 25 to 200 mg/kg (Whitehead, 1966). Forage values ranged from 44 to 62 mg/kg at several sites in N e w Brunswick (Nicho lson , 1982) and from 39 to 351 mg/kg at several sites in N e w Zealand. M n content is considerably inf luenced by so i l factors, especial ly p H and drainage. Wel l -d ra ined , calcareous soils were associated with low M n levels ( M i t c h e l l , 1964) and l iming reduces M n levels (Mi tche l l , 1964; Reith and Mi t che l l , 1964; Metson et al . , 1979). 4.6.10.2 Agassiz : M n contents were s l igh t ly higher in 1985 than in 1984. N o differences between L I and H F cutt ing were evident but there were cul t ivar differences with the order as fo l lows: Pra i r ia l (124) > Barlano (122) > Nor lea (114) > Sumas (103). However, these differences are not at 178 all similar to those of Mitche l l (1964) who found that orchardgrass contained two to three times as much M n as ryegrass. Table 44: Average Manganese Content of Herbage at Agassiz. 1984 &1985 Cul t ivar Cutting Year W e i g h t e d T r e a t m e n t 1984 1985 Mean * M n Content (mg/kg) Pra ir ia l L I 128 125 127 Pra i r ia l H F 114 129 122 Sumas L I 94 115 104 Sumas H F 88 116 102 Barlano L I 113 137 124 Barlano H F 105 135 120 Nor lea L I 106 1 24 114 Nor lea H F 104 124 114 Weighted Mean * 10 6 126 116 *Notc: Weighted according to number of samples analyzed. Seasonal changes were evident with herbage M n levels steadily increasing • from the first cut lows (mean, 86) to highs by late September/early October ( mean, 145mg/kg) with a subsequent drop in late-October (except for H F cutting in 1985). This seasonal pattern of M n content agrees with that found by Hemingway (1962) who noted an i n c r e a s e f r o m l a t e - M a y to m i d - S e p t e m b e r in the U . K . TABLE 45: MANGANESE CONTENT FOR OYSTER RIVER FERTILIZATION TRIAL. 1985 FERTILIZATION HARVEST DATE YEARLY REGIME MAY 9 JUNE 10 JULY 8 AUG 12 SEPT 11 OCT 29 MEAN MAY 2 1986 MANGANESE CONTENT (mg/kg) 1985 I 47 63 96 86 129 143 94 67 II 51 71 82 100 141 160 101 79 III 47 68 86 91 124 136 92 75 IV 45 58 82 96 114 112 85 77 HARVEST MEAN 48 65 87 93 127 138 75 180 4.6.10.3 Ovslcr River: Average M n content differed between the fertilizer regimes with the following ranking: II (101) > I (94) > I I I (92) > I V (85). No apparent differences due to clover content are evident. The literature reports conflicting evidence on M n content of clover relative to grass. Fleming (1965) reported that clovers can be lower or higher than grasses in M n content. Others have found clover to have higher M n levels than grasses (Whitehead, 1966) whereas the reverse has also been recorded (Hemingway, 1962; Reay and Marsh, 1976; Metson et al, 1979). No consistent fertilizer regime effects were evident. N application may increase M n levels but the effect is related to the effect of fertilizer on p H and its subsequent influence on M n levels (Whitehead, 1966). The distribution of yield was similar to that at Agassiz with a steady increase from early May (48) to late October (138mg/kg). No residual effects were evident from regime I V on May 2, 1986. The grass plots had higher M n levels (mean, 77) than the high clover plot (regime I, 67mg/kg), a result that could be related to relative clover content. 4.7 . D I S C M E T E R I N V E S T I G A T I O N S The results from the measurements conducted on pure plots of Prairial orchardgrass and Barlano perennial ryegrass show that there were significant relationships (P < 0.001) between the height of herbage and D M yield (Table 46). The percentage of variation explained by the regression ( r 2 values) varied from 38.4% to 82.6%. Results from measurements taken under grazing situations showed a similar significant (P < 0.01) relationship 181 between herbage height and D M yields (r 2 values ranged from 34.2% to 90.1%). In general the range of D M yields was greater for measurements conducted under grazing situations and the r values were lower (Table 46) due to the variability of the grazed pastures. Although the instrument cannot be claimed to be highly accurate (since standard errors, of the estimates are high), it does have several advantages. The disc instrument was inexpensive (approximately $12.00 for materials), easy to construct and readily repaired. Phillips and Clark (1971) noted that the instrument can be easily duplicated and requires little skill to operate. In addition, an operator can be trained in a short time and it can be used quickly (e.g. 50 readings over a 2.5 ha. field in 15 minutes) . The instrument used in these investigations was based on the design of Castle (1976) who obtained highly significant relationships between height and D M yield. For pure grass swards r 2 values ranged from 85.2 to 90.1%. Under grazing situations the r 2 values were 61.5% and 38.7%. This greater accuracy obtained by Castle (1976) could in part be due to the method of calibration. Castle (1976) measured disc height at three different positions on pure grass plots (3.66 X 1.52 m). The relationship between average height and plot yield was then determined by harvesting the plot with an autoscythe, weighing the total sample and determining D M of a subsample. Under grazing conditions, disc height measurements were made at five postions within each of five sites which were then harvested with an autoscythe. No calibration cuts were made to soil level as was done in this experiment. Sampling to soil level wil l cause greater variability in the measurements and in general lower r values. Cuts taken to ground level contain greater portions of stubble which do not 182 necessarily affect disc meter height. Sampling with an autoscythe collects mostly leaves which is better correlated to disc height than herbage harvested to soil level to include leaves and stubble. Improved regressions can undoubtedly be obtained under local conditions by harvesting calibration cuts above soil level (i.e. 3 to 6 cm). DISC HEIGHT (cm) 183 In can be seen that there are marked differences in the regressions between the four types of herbage sampled (Table 46). The slopes (b) for Prairial orchardgrass were smaller than those for Barlano ryegrass; a result also observed by Castle (1976). The Langley pasture site was predominantly orchardgrass and had slopes very similar to those of Prairial orchardgrass. Slopes at Farm #2 showed large differences due to the variability of the herbage (grass and red clover) and the stemminess of the red clover. In general there was a change in regression from Apri l to July. The slopes tended to decline from Apri l to July indicating differences in regression for different times of the year. Phillips and Clark (1971), Powell (1974) and Castle (1976) also noted changes in regression with different seasons. Poor regressions were evident at each of the 4 sites at different times of the year. The regressions for Prairial on June 25 and for Barlano on July 14 showed poor relationships between D M yield and disc height. Herbage at these times generally showed a large amount of inflorescence development. The more stemmy material tended to hold the disc, not allowing it to settle. The same problem occurred under grazing situations during A p r i l and M a y where herbage growth had exceeded cattle intake and the grass had become somewhat mature. The herbage at Langley had developed to the seed head stage in early May which is reflected by the poor correlation between D M yield and herbage height on May 13. Regressions at Farm #2 also showed lower r values during April and mid-June due to red clover which had a very stemmy growth form. TABLE 46 : THE RELATIONSHIP BETWEEN THE YIELD OF DRY MATTER MOOkq/ha) AND THE HEIGHT OF THE HERBAGE (cm). 1986 LOCATION HERBAGE DATE NUMBER OF YIELD OF D.M. CORRELATION COEFFICIENTS OF OBSERVATIONS MIN. MEAN • MAX. COEFFICIENT REGRESSION r EQUATION a b AGASSIZ PRAIRIAL APRIL 10 6 11.60 24.82 44.84 0.899 * 6.20 2.24 PLOTS MAY 14 20 14.71 25.52 52.19 0.770 ... 8.69 0.94 MAY 23 20 15.42 23.59 32.39 0.774 ... -3.70 1.50 JUNE 23 10 17.54 26.24 36.35 0.854 * * 3.13 1.41 JUNE 25 10 19.24 26.93 35.22 0.620 7.87 1.15 AGASSIZ BARLANO APRIL 15 10 13.58 25.50 34.09 0.866 5.60 2.50 PLOTS MAY 14 10 16.41 35.90 51.20 0.909 * * * -14.13 2.38 MAY 23 20 17.40 34.88 55.16 0.795 * * * -8.13 2.06 JUNE 25 10 24.61 37.37 67.75 0.853 -34.00 4.53 JULY 14 20 18.67 26.71 38.76 0.629 * * 5.55 1.34 AGASSIZ GRASS / APRIL 10 10 15.13 2.8.13 49.50 0.759 • -2.10 3.39 FARM #2 CLOVER APRIL 30 10 22.77 32.80 47.38 0.728 * 21.30 1.10 PASTURE JUNE 18 10 14.57 21.30 28.71 0.745 * 9.86 0.93 JUNE 27 10 18.25 32.65 60.40 0.949 # * * -5.33 3.05 LANGLEY GRASS / MAY 13 10 23.76 34.55 41.73 0.585 11.88 1.75 CLOVER MAY 16# 10 19.94 29.86 41.87 0.740 * 12.10 1.03 PASTURE MAY 16# 10 14.43 23.65 37.06 0.860 10.00 1.18 JULY 3 20 13.30 23.28 34.65 0.727 • » * 7.59 1.37 #NOTE: First measurements made on pre-graze and second on post graze pasture. * P < 0.05 ** P < 0.01 *** P < 0.001 oc 185 The good regression at Farm #2 on June 27 (r = 0.949) was measured on a predominantly grass sward. The disc meter proved to be a very easy to use device for measuring D M yield under both plot and grazing situations. As has been shown, the meter does not have a high degree of accuracy but it can be used for some very useful non-destructive measurements. The disc meter would be very useful in measuring yields of uninterrupted growth and can also be used to determine when herbage growth starts in the spring (Jagtenberg, 1970). For grazing experiments, the instrument can be used, to determine when animals should be moved on the basis of pre- or post-grazing heights and once animals are grazing it can be used to record the amount of herbage on offer. However, before the device can be utilized locally it is necessary to conduct further studies in order to calibrate it for various varieties, seasons and herbage conditions (i.e. grazed pastures, pure plots and grass/clover mixtures). Once calibrations have been developed, the disc meter would be very useful for estimating the yield of experimental plots. Based on personal experience, the disc meter is also very useful as a farm management aid in determining feed on offer, residual dry matters and pre-grazing herbage yields. The disc meter has been used frequently in the U . K . and in New Zealand for both experimental and practical farm work (Phillips and Clark, 1971; Powell, 1974; Castle and Watson, 1975 and Castle, 1976). Recently the 'sward stick' has become a popular research and management aid in the U . K . since most pastures consist of dense perennial ryegrass swards where 'sward stick height' (SSH) is well correlated to herbage D M yield. 186 V C O N C L U S I O N S The seasonal distribution of yield and total annual yield of herbage are obviously affected by climatic factors, but it is also evident that yields can be greatly affected by management factors such as variety selection, fertilization, defoliation regimes and irrigation. Over the two years of investigations at Agassiz , orchardgrass (Prairial and Sumas) produced greater annual yields than perennial ryegrass (Barlano and Norlea). Under low-infrequent (LI) cutting, perennial ryegrass produced a larger proportion of annual yield during the spring compared to orchardgrass which produced more of its annual yield during the summer. This distribution can be altered under high-frequent (HF) cutting which essentially shifts some spring production into the summer months. H F cutting produced greater yields during the summer than L I cutting, especially during periods of low rainfall. Although seasonal yields were affected by cutting regime, annual yields showed very little difference. Annual yield varied largely due to differences in climate between years and with cultivars within years. From a management viewpoint these results agree with general grazing management recommendations (Hal l , 1973). During the spring, when growth rates are high, defoliation should be more severe to provide better herbage production and greater efficiency of utilization. During the summer, and especially during dry periods, grazing should be more lenient. Traditional set stocking does not offer the degree of control required to achieve these grazing intensities. The use of rotational or buffer grazing systems would be necessary to achieve the desired levels of animal control. 187 The trial at Oyster River demonstrated that the distribution of yield can be affected by split N applications. Effects of irrigation were evident during the dry summer period. Appl icat ion of N during August significantly increased fall yields which can be very useful in providing an extended grazing period. The same fertilization treatment also resulted in higher spring yields for the next year. Knowledge of seasonal growth patterns for specific cultivars and how these can be affected by fertilization regimes would be very useful in farm management and feed budgeting. Certain areas can be planted to selected varieties or mixtures to better balance feed production with animal requirements. In addition, fertilization regime can be varied to further improve the balance between feed production and requirements resulting in improved herbage utilization. Dairy, beef and sheep enterprises could benefit greatly from feed budgeting strategies for herbage. In certain cases, concentrate supplementation may be reduced, but the most important benefit wil l occur as a result of more efficient allocation of concentrates. Feed budgeting will allow for better farm planning and extend the grazing season through summer dry periods and further into the fall. Simple measurement devices are essential for improved grassland management. The disc meter was shown to demonstrate a good relationship between herbage yield and height under certain conditions. Disc meters and similar instruments such as the 'sward stick' are inexpensive and simple to use and further investigation of such devices is w a r r a n t e d . VI SUGGESTIONS FOR FURTHER RESEARCH 1. Examine seasonal distribution of yield and annual yield of various herbage varieties in pure swards and mixtures. 2. Evaluate varieties for pasture and mixed management situations by both cutting methods (i.e. method of Anslow and Green, 1967) and grazing experiments (Orr et al., 1988). 3. Further examination of seasonal distribution of growth as affected by split fertilization regimes. 4. To determine if fall application of N can give consistent and economic yield responses in the spring of the following year and to examine root mass and root reserves which may be related to this yield response. 5. Examination of yield response from summer irrigation. 6. Further evaluation of disc meter and sward stick for use in experimental plot work and practical farm management. 189 B I B L I O G R A P H Y Adams, A . F . R . and Elphick, B . L . (1956). The copper content of some soils and pasture species in Canterbury. New Zealand Journal of Science and Technology A38: 345-358. Ahlgren, H . L . ; Sund, J . M . ; Zehner, C . E . ; Allen, N . N . and Graul, E . J . (1951). Comparisons of the productivity of permanent and rotation pastures on plowable cropland. Part I. Agronomy Journal. Vol.43. 500-503. Ahloowalia, B.S.; Brogan, J . C . and Sherington, J. (1981). Yields and quality performance of populations derived from tetraploid Italian ryegrass. Irish Journal of Agricultural Research, Vol .20, 191-200. Alberda, T h . (1960). The effect of nitrate nutrition on carbohydrate content in Lolium perennc L . Proceedings of the VIII International Grassland Congress (Reading, England). 612-615. Anderson, L . B . ; Brock, J .L . ; Boyd, A . F . ; Harris, A . J . and Ryan, D . L . (1982). 'Grasslands Roa' tall fescue: herbage dry matter production and quality under mowing. New Zealand Journal of Experimental Agriculture. V o l . 10 275-280. Anslow, R . C . (1965). Grass growth in midsummer. Journal of the British Grassland Society. V o l . 20, 19-26. Anslow, R . C . (1967). Frequency of cutting and sward production. Journal of Agricultural Science, Cambridge. V o l . 68, 377-384. Anslow, R . C . and Back, I L L . (1967). Grass growth in midsummer and light interception and growth rate of a perennial ryegrass sward. A reply to a recent re-interpretation of published data. Journal of the British Grassland Society. V o l . 22, 108-111. Anslow, R . C . and Green, J .O. (1967). The seasonal growth of pasture grasses. Journal of Agricultural Science, Cambridge. V o l . 68, 109-122. Appadurai, R.R and Holmes, W. (1964). The influence of stage of growth, closeness of defoliation and moisture on the growth and productivity of a ryegrass - white clover sward. I. Effect on herbage yields. J. Agric. Sci., Camb. 62 : 327-332. 190 B . C . M . A . & F . (1983). Mineral Analysis Results. South Coastal Forage Competition, 1983. B . C . Ministry of Agriculture and Food. B . C . M . A . & F . (1985). Average analysis of British Columbia Feeds. 1969-1985. District: Vancouver Island. B . C . Ministry of Agriculture and Food. 19-21. B . C . M . A . & F . (1986). Mineral Analysis Results. South Coastal Forage Competition, 1983. B . C . Ministry of Agriculture and Food. Baars, J . A . (1976). Seasonal distribution of pasture production in New Zealand. IX. Hamilton. New Zealand Journal of Experimental Agriculture. Vol .4 , 157-161. Baars, J . A . (1976a). Seasonal distribution of pasture production in New Zealand. VIII Dargaville. New Zealand Journal of Experimental Agriculture. Vol .4 . 151-156. Baars, J . A . (1976b). Seasonal distribution of pasture production in New Zealand. IX Hamilton. New Zealand Journal of Experimental Agriculture. Vol .4 . 157-161 Baars, J .A . : Radcliffe, J . E . and Brunswick, L . (1975). Seasonal distribution of pasture production in New Zealand. VI Wairakci, pasture and lucerne production. New Zealand Journal of Experimental Agriculture. Vol .3 . 253-258. Bakhuis, J . A . (1960). Estimating pasture production by use of grass length and sward density. Netherlands Journal of Agricultural Science. V o l . 8, 211-224. Bales, D . (1986). Grass varieties for 1986. Field Crop Facts, South Coastal Forage Notes. Province of British Columbia, Ministry of Agriculture and Food. 3 pp. Bcever, D .E . ; Terry, R .A . ; Cammell, S.B.. and Wallace, A.S . (1978). The digestion of spring and autumn harvested perennial ryegrass by sheep. Journal of Agricultural Science, Cambridge. Vol.90, 463-470. Black, W . J . M . and Alexander, J . R . B . (1967). The effects of type of cutter and height of cutting on the recovery of four grass varieties. Journal of the British Grassland Society. V o l . 22, 260-263. 191 Bland, B . F . (1967). The effect of cutting frequency and root segregation on the yield from perennial ryegrass - while clover associations. J. Agric. Sci., Camb. 69: 391-397. Bomke, A . A . and Bertrand, R . A . (1979). Forage response to rate and method of urea and ammonium nitrate application. Province of British Columbia, Ministry of Agriculture and Food. Soils Note. 2pp. Bosch, S. (1956). The determination of pasture yield. Netherlands Journal of Agricultural Science. Vol.4. 305-313. Boswell, C C . (1980). Pasture production: measurement techniques. New Zealand Ministry of Agriculture and Food. Aglink Series, A S T 86. Boswell, C C . and Mosgiel, P .B. (1977). Effects of cutting regime on pasture production. New Zealand Journal of Experimental Agriculture. V o l . 5, 4 0 3 - 4 0 8 Boyd, D . A . (1949). Experiments with leys and permanent grass. Journal of the British Grassland Society. Vol.4. 1-10. Brelin, B . (1979) Mixed grazing with sheep and cattle compared with single grazing. Swedish Journal of Agricultural Research. V o l . 24, 2 2 1 - 2 2 9 . Brockman, J.S. (1966). The growth rate of grass as influenced by fertilizer nitrogen and stage of defoliation. Proceedings of the X International Grassland Congress (Helsinki, Finland). 234-240. Brougham, R . W . (1955). A study in rate of pasture growth. Australian Journal of Agricultural Research. V o l . 6, 804-812. Brougham, R . W . (1956a). The rate of growth of short rotation ryegrass pastures during the late autumn, winter and early spring. New Zealand Journal of Science and Technology. Section A . V o l . 38, 78-87. Brougham, R . W . (1956b). Effect of intensity of defoliation on regrowth of pasture. Australian Journal of Agricultural Research. V o l . 7, 377-387. 192 Brougham, R . W . (1959). The effects of frequency and intensity of grazing on the productivity of a pasture of short-rotation ryegrass and red and white clover. New Zealand Journal of Agricultural Research. V o l . 2, 1232-1248. Brougham, R . W . (1959). The effects of season and weather on the growth rate of a ryegrass and clover pasture. New Zealand Journal of Agricultural Research. V o l . 2, 283-296. Brougham, R . W . and Glenday, A . C . (1967). Grass growth in midsummer - a reinterpretation of published data. Journal of the British Grassland Society. V o l . 22, 100-107. Brown, B . A . and Munsell, R.I. (1945). Deterioration of clipped caged areas in permanent pastures. Journal of the American Society of Agronomy. Vol.37. 542-548. Brown, D . (1954). Methods of surveying and measuring vegetation. Commonwealth Bureau of Pastures and Field Crops. Hurley, Berkshire, England. Bulletin #42. C . A . B . Bryan, W . W . and Sharpe, J.P. (1965). The effect of urea and cutting treatments on the production of Pangola grass in South East Queensland. Australian Journal of Experimental Agriculture and Animal Husbandry. V o l . 5, 433-441. Burger, A . W . ; Kackobs, J .A . and Hittle, C . N . (1958). The effect of height and frequency of cutting on the yield and botanical composition of tall fescue and smooth bromegrass mixtures. Agronomy Journal. V o l . 50, 629-632 . Calder, D . M . and Davies, A . (1965) Spring growth of perennial ryegrasses and cocksfoots. Journal of the British Grassland Society. Vo l . 20, 194-196. Cameron, I .H. and McGowan, A . A . (1969). Responses by pastures to current and previous applications of superphosphate. Australian Journal of Experimental Agriculture and Animal Husbandry. Vol .9 , 617-621. Campbell , A . G . (1966). Grazed pasture parameters. I. Pasture dry matter production and availability in a stocking rate and grazing management experiment with dairy cows. Journal of Agricultural Science. Vol.67. 199-210. 193 Curler, J .F. (1962). Herbage sampling for yield: tame pastures. In: Pasture and Range Research Techniques. Edited by American Society of Agronomy and others. Comstock publishing associates, Ithaca, U.S.A. Castle, M . E . (1976). A simple disc instrument for estimating herbage yield. Journal of the British Grassland Society. Vol.31, 37-40. Castle, M . E . and Watson, J .N . (1975). Further comparisons between a rigid rotational "Wye College' system and other systems of grazing for milk production. Journal of the British Grassland Society, Vol.30, 1-6. Castle, M . E . ; Reid, D . and Hcddle, R . G . (1965). The effect of varying the date of application of fertilizer nitrogen on the yield and seasonal productivity of grassland. Journal of Agricultural Science, Cambridge. V o l . 64, 177-184. Chestnutt, D . M . B . and Lowe, J. (1970). Agronomy of white clover/grass swards: review. In: Lowe, J. (Ed.) White clover. Occasional Symposium of the British Grassland Society. No. 6, 191-213. Christian, G . D . and Feldman, F .J . (1970). Atomic Absorption Spectroscopy. Applications in Agriculture, Biology and Medicine. Pub. Willey-Interscience. New York, London, Sydney, Toronto. Clancy, M . E . and Wilson, R . K . (1966). Development and application of a new chemical method for predicting the digestibility and intake of herbage samples. Proceedings of the 10th. International Grassland Congress, (Helsinki, Finland). 445-453. Cooke, D . A . (1969). Estimating herbage from simulated grazing trials. In: Baden Campbell, J. (comp.). Experimental Methods for Evaluating Herbage. Canada Department of Agriculture. Publication 1315, 80-90. Corral I, A . J . (1978). The effect of genotype and water supply on the seasonal pattern of grass production. 7th General Meeting of the European Grassland Federation (Gent) 2.33-2.32. 194 Corrall, A . J . and Fenlon, J.S. (1978). A comparative method for describing the seasonal distribution of production from grasses. Journal of Agricultural Science, Cambridge. Vol.91, 61-67. ' Cowling, D . W . and Lockyer, D.R. (1965). A comparison of the reaction of different grass species to fertilizer nitrogen and to growth in association with white clover. I. Yield of dry matter. Journal of the British Grassland Society. Vol.20, 197-204. Culleton, N . and Fleming, G . A . (1983). Mineral composition of ryegrass cultivars. Irish Journal of Agricultural Research. Vol.22, 21-29. Currier, C . G . ; Haaland, R . L . ; Hoveland, C.S.; Elkins, C B . and Odom, J.W. (1981). Breeding for higher magnesium content in orchardgrass (Dactylis glomerala L . ) . Proceedings of the X I V International Grassland Congress, Lexington, U . S . A . 127-129. Davies, J . G . and Trumble, H . C . (1934). Notes on the technique of pasture investigations. Imperial Bureau of Plant Genetics: Herbage Plants. Bulletin #14. Grassland Research in Australia. 23-32. Davis, W . E . P . (1960). Effect of clipping at various heights on characteristics of regrowth in reed canary grass (Phalaris arundinacea L . ) . Canadian Journal of Plant Science. Vol.40, 452-456. Davis, W . E . P . (1961). An evaluation of varietal characteristics of orchardgrass (Dactylis glomerala L.) subjected to a silage-pasture type of utilization. Canadian Journal of Plant Science. Vol.41, 653-663. Deinum, B . and Sibma, L . (1980). Nitrate content of herbage in relation to nitrogen fertilization and management. In: Prins, W . H . and Arnold, G . H . (Editors). The role of nitrogen in intensive grassland production. Pub. Pudoc. Wageningen, the Netherlands. Dickson, I.A.; Frame, J. and Arnot, D.P. (1981). Mixed grazing of cattle and sheep versus cattle only in an intensive grassland system. Animal Production. Vol.33. 265-272. Di lz , K . (1966). The effect of nitrogen nutrition and clipping frequency on regrowth of perennial ryegrass. Proceedings of the X International Grassland Congress (Helsinki). 160-164. 195 Dowdell, R.J.; Morrison, J. and Hood, A . E . M . (1980). The faye of fertilizer nitrogen applied to grassland: uptake by plants, immobilization into soil organic matter and losses by leaching and denitrification. In: W . H . Prins and G . H . Arnold (Eds.), The Role of Nitrogen in Intensive Grassland Production. Pudoc, Wageningen, the Netherlands. 129-136. DuChatcau, P .C. ; Nafziger, D . L . ; Ahuja, L . R . and Swartzenruber, D (1972). Experimental curves and rates of change from piecewise parabolic fits. Agronomy Journal. V o l . 64. 538-544. Earle, D . F . and McGowan, A . A . (1979). Evaluation and calibration of an automated rising plate meter for estimating dry matter yield of pasture. Australian Journal of Experimental Agriculture and Animal Husbandry. Vol .19, 337-343. Edye, L . A . (1969). Yield comparisons of thirty eight introductions of Glycine javanica in swards in three environments. Australian Journal of Experimental Agriculture and Animal Husbandry. V o l . 7, 342 -350 . Elliott, L L . and Lynch, P.B. (1958). Techniques of measuring pasture production in fertilizer trials. New Zealand Journal of Agricultural Research. Vol .1 . 498-521. Fairbourn, M . L . (1983). First harvest phonological stage effects on production of eight irrigated grasses. Agronomy Journal. Vol.75, 189-192. Fairey, N . A . (1985a). Productivity and quality of perennial and hybrid ryegrass, orchardgrass and reed canarygrass grown in the lower mainland of British Columbia. Canadian Journal of Plant Science. Vol .65 , 1 17-124. Fairey, N . A . (1985b). Productivity, quality and persistence of perennial ryegrass as influenced by cutting/fertility management and ploidy. Canadian Journal of Plant Science. Vol.65, 565-571. Featherstone, J . ; Rickarby, C D . and Cavcll , A . J . (1951). Variations in the progressive yield and chemical composition of grassland herbage during a period of three months growth. Journal of the British Grassland Society. V o l . 6, 161-166. 196 Fleming, G . A . (1965). Trace elements in plants with particular reference to species. Outl. Agric. Vol.4. 270-285 (cited from Whitehead, 1966). Fleming, G . A . and Coulter, B.S. (1963). Mineral elements in pasture plants. Proc. 1st Reg. Conf.. Int. Potash Inst... Wexford (Ireland) 1963, 63-70. (Cited from Whitehead, 1966). Frame, J. (1965). The effect of cutting and grazing techniques on productivity of grass-clover swards. Proceedings of the IX International Grassland Congress (Sao Parlo) 1511-1516. Frame, J. (1966). The evaluation of herbage production under cutting and grazing regimes. Proceedings of the X International Grassland Congress (Helsinki) 291-297. Frame, J. (1981). Herbage mass. In: Sward Measurement Handbook. Edited by J. Hodgson et al. Published by the British Grassland Society. 39-69. Frame, J. (1984). Written communication. Frame, J . and Boyd, A . G . (1986a). Effect of cultivar and seed rate of perennial ryegrass and strategic fertilizer nitrogen on the productivity of grass / white clover swards. Grass and Forage Science, V o l . 41, 359-366. Frame, J. and Boyd, A . G . (1986b) Response of white clover to strategic fertilizer nitrogen in spring and/or autumn applied to a grass / white clover sward. In: Cooper, J.P. and Raymond, W . F . (Eds) Grassland Manuring. Occasional Symposium of the British Grassland Society No. 20, Frame, J. and Boyd, A . G . (1987a). The effect of fertilizer nitrogen rate, white clover variety and closeness of cutting on herbage productivity from perennial ryegrass / white clover swards. Grass and Forage Science, Vol.42, 85-96. Frame, J. and Boyd, A . G . (1987b). The effect of strategic use of fertilizer nitrogen in spring and/or autumn on the productivity of a perennial Fuelleman, R . F . and Burlison, W . L . (1939). Pasture yields and consumption under grazing-conditions. Journal of the American Society of Agronomy. Vol.31, 399-412. 197 Fuelleman, R . F . and Burlison, W . L . (1940). A comparison of yields and composition of some Illinois pasture plants. Journal of the American Society of Agronomy. V o l . 32, 243-255. Gardner, A . L . and Alberquerque, II. (1965). Seasonal growth of various forage species in Uruguay. Proceedings of the IX International Grassland Congress. V o l . 2, 1053-1058. Garwood, E . A . ; Salette, J. and Lemaire, G . (1980). The influence of water supply to grass on the response to fertilizer nitrogen and nitrogen recovery. In: Prins, W . H . and Arnold, G . H . (Editors). The role of nitrogen in intensive grassland production. Pub. Pudoc, Wegeningen, the Netherlands. Glenday, A . C . (1955). The mathematical separation of plant and weather effects in field growth studies. Australian Journal of Agricultural Research. V o l . 6, 813-822. Glenday, A . C . (1959). Mathematical analysis of growth curves replicated in time. New Zealand Journal of Agricultural Research. V o l . 2, 297-305. Goodenough, D . C . W . ; MacDonald, C.I. and Morrison, A . R . J . (1984). Growth patterns of Italian Ryegrass cultivars established in different seasons. Journal of the Grassland Society of South Africa. Vo l . 1, 21-24. Grassland Research Institute. (1961). Research techniques in use at the Grassland Research Institute, Hurley. Commonwealth Bureau of Pastures and Field Crops, Hurley, Berkshire, England. Bulletin #45. Green, J .O. (1959). The measurement of herbage production. In: The Measurement of Grassland Productivity. Edited by J .D. Ivins. Butterworths Scientific Publishers, London. Griffith, G . ap (1960). Nitrate content of herbage at different manurial levels. Nature. Vo l . 185, 627-628. Griffith, G . ap and Wallers, R .J .K. (1966). The sodium and potassium content of some grass genera, species and varieties. J. agric. Sci. Camb. V o l . 67, 81-89. Griffiths, G . L . (1958). Nitrate content of some grass species and strains. Nature. V o l . 182, 1099. 198 Grunes, D . L . (1967). Grass Tetany. In: McGraw-Hil l Yearbook Science and Technology. Pub. McGraw-Hi l l Book Co. 191-193. Hal l , A . M . (1973). The principles of pasture growth. Dairyfarming Annual. 102-105. Havre, G . M . and Dishington, I.E. (1962). The mineral composition of pasture as influenced by various types of heavy nitrogen dressings. Acta Agric. Scand. V o l . 12, 298-308. Hayman, J . M . and Moss, R . A . (1979). Grazing irrigated pasture with sheep: effect of winter management on pasture regrowth. New Zealand Journal of Experimental Agriculture. Vol .7 , 277-280. Heard, C . A . H . and Wiseman, I.G.R. (1973). Use of annual production curves of pasture species in pasture planning. Proceedings of the Grassland Society of South Africa. Vol.8, 43-46. Heckman, M . (1967). Minerals in feeds by atomic absorption spectrophotometry. J . A . O . A . C . V o l . 50, 45-50. Hemingway, R . G . (1961). Magnesium, potassium, sodium and calcium contents of herbage as influenced by fertilizer treatments over a three year period. Journal of the British Grassland Society. Vol.16, 106-116. Hemingway, R . G . (1962). Copper, molybdenum, manganese and iron contents of herbage as influenced by fertiliser treatments over a three year period. Journal of the British Grassland Society. Vol.17, 1 8 2 71 8 7. Henkens, C H . (1965). Factors influencing the sodium content of meadow grass. Netherlands Journal of Agricultural Science. Vol.13, 21-47. Henzell, E . F . and Stirk, G . B . (1963). Effects of nitrogen deficiency and soil moisture stress on the growth of pasture grasses at Samford, South East Queensland. 1. Results of field experiments. Australian Journal of Agricultural Research and Animal Husbandry. V o l . 3, 300-306. Herbage Plants. Bulletin #11. Techniques Employed in Grassland Research in New Zealand. September, 1933. 21-35 199 Hodgson, J.; Baker, R.D.; Davies, A . ; Laidlaw, A . S . and Leaver, J .D. (Editors). (1981). Sward Measurement Handbook. Pub. The British Grassland Society. 277pp. Holmes, C . W . (1974). The Massey grass meter. New Zealand Dairy Farming Annual. p27. Holmes, W.; Waite, R.; Fergusson, D . L . and Campbell, J.L (1950). Studies in grazing management. I. A comparison of the production obtained from close-folding and rotational grazing of dairy cows. Journal of Agricultural Science. V o l . 40, 381-391. Hudson, A . W . (1933). Measurement of pasture production: Extension and demonstration fricld trials. Imperial Bureau of Plant Genetics: Jaglcnburg, W . D . (1970). Predicting the best time to apply nitrogen to grassland in spring. Journal of the British Grassland Society. Vol.25, 2 6 6 - 2 7 1 . Janti, A . and Kramer, P.J. (1956). Regrowth of pastures in relation to soil moisture and defoliation. Proceedings of the VII International Grassland Congress. 33-42. Johns, A . T . (1955). Pasture quality and ruminant digestion. I. Seasonal change in botanical and chemical compostion of pasture. New Zealand Journal of Science and Technology. Section A , Vol.37, No.4, 301 -311 . Johnson, A . D . (1978). Sample preparation and chemical analysis of vegetation. In: Measurement of Grassland Vegetation and Animal Production. E d . L . 't Mannetje. Pub. Commonwealth Agricultural Bureaux, Farnham Royal, pp 96. Joint Committee of the American Society of Agronomy, American Dairy Science Association and American Society of Animal Production. (1943). Preliminary report on pasture investigation techniques. Journal of Dairy Science. Vol.26 (4), 353-369. Jones, D .I .H. (1963). The mineral content of six grasses from a Hyparrhenia-dominant Grassland in Northern Rhodesia. The Rhodesian Journal of Agricultural Research. V o l . 1 (1). 35-38. 200 Jones, R.I. and Bartholomew, P .E . (1973). Production curves for pastures and Veld. Proceedings of the Grassland Society of South Africa. Vol .8 , 47-52. Kaltofen, H . ; Kreil , W.; Kasdorff, K . and Leistner, J. (1966). The effects of heavy doses of nitrogen applied to pasture in spring compared with split applications given during the vegetative period. Proceedings of the X International Grassland Congress (Helsinki, Finland). 231-234. Kean, G . D . (1982). The annual yield, and its distribution, of some grass cultivars and mixtures. Irish Journal of Agricultural Research. Vol .21 . 159-169. Kemp, A . (1960). Hypomagnesemia in mikling cows: The response of serum magnesium to alterations in herbage composition resulting from potash and nitrogen application on pasture. Netherlands J. agric. Sci. V o l . 8, 281-304. Kemp, O.R. (1973). The seasonal growth of tropical pasture grasses on the mid-North Coast of New South Wales. M.Sc . Agr. Thesis. University of Sydney, Australia. 309pp. Knight, R. (1968). The seasonal growth rhythm of some cultivars of cocksfoot (Dactylis glomerta L . ) . Australian Journal of Agricultural Research and Animal Husbandry. V o l . 8, 309-316. Laidlaw, A . S . (1980) The effects of nitrogen fertilizer applied i n spring in swards of ryegrass sown with four cultivars of white clover. Grass and Forage Science, 35, 295-299. Laidlaw, A . S . (1984) Quantifying the effect of nitrogen fertilizer applications in spring on white clover content in perennial ryegrass-white clover swards. Grass and Forage Science, 39, 317-321. Laissus, R. (1983) How to use nitrogen fertilizers on a grass-white clover stand. In: Corrall, A . J . (Ed) Efficient Grassland Farming. Proceedings of the 9th General Meeting of the European Grassland Federation published as Occasional Symposium of the British Grassland Society. No. 14, 223-226. 201 Lambert, D . A . (1962). A study of growth in swards of timothy and meadow fescue. III> The effect of two levels of nitrogen under two cutting treatments. Journal of Agricultural Science, Cambridge. Vol .59 , 25-32. Lampeter, W . (1960). Mutual influences of higher plants regarding growth of shoots and roots, mineral content and water consumption studied in some economically important forage plants . Wiss. Z . Karl-Marx-Univ. Leibzig. 1959-1960 Vol.9 611-722. (Cited in Herbage Abstracts Vol.32, 2061). Langer, R . H . M . (1958). A study of growth in swards of timothy and meadow fescue. I. Uninterrupted growth. Journal of Agricultural Science, Cambridge, V o l . 51, 347-352. Langer, R . H . M . (1959). A study of growth in swards of timothy and meadow fescue. II. The effects of cutting treatments. Journal of Agricultural Science, Cambridge, V o l . 52, 273-281 Large, R . V . ; Cobby, J . M . and Baker, R . D . (1985). The design and conduct of experiments to measure animal and herbage production responses to fertilizer nitrogen under cutting and grazing managements. Journal of Agricultural Science, Cambridge, Vol.104, 85-94. Leaver, J .D. (1982) Editor. Herbage Intake Handbook. Pub. by the British Grassland Society, Hurley, U . K . Lightncr, J.W.; Rhykerd, C . L . ; Van Scoyoc, G . E . ; Men gel, D .B . ; Noller, C H . and Blair, B . O . (1981). Effect of N - P - K fertilization on mineral composition of Poa protensis L . grown on a shallow muck soil. Proceedings of the X I V International Grassland Congress (Lexington, U.S .A. ) 294-296. Linehan, P .A . ; Lowe, J. and Stewart, R . H . (1952). The output of pasture and its measurement. Part III. Journal of the British Grassland Society. Vol .7 , 73-98. Lush, R . H . (1938). Seasonal composition and yields of pastures. Louisiana State University and Agricultural and Mechanical College. Agricultural Experiment Stations. Louisiana Bulletin No.299. 23pp. 202 Lynch, P .B . (1947). Methods of measuring the production from grasslands. A review of the techniques employed by the fields division, Department of Agriculture. New Zealand Journal of Science and Technology. Section A . Vol.28. 385-405. Lynch, P .B . and Mountier, N.S. (1954) Cutting techniques in grassland experiments. New Zealand Journal of Science and Technology. Vo l .36A, 375-385. MacLeod, L . B . (1965). Effect of nitrogen and potassium on the yield and chemical composition of alfalfa, bromegrass, orchardgrass and timothy grown as pure species. Agronomy Journal. Vol.57 261-266. MacLeod, L . B . ; Bishop, R.F. ; Calder, F .W. and MacEachern, C R . (1964). Effects of various rates of liming and fertilization on certain chemical properties of a strongly acid soil and on the establishment, yield, botanical and chemical composition of a forage mixture. Canadian Journal of Soil Science. Vol.44, 237-247. 't Mannetje, L . (1978). Measuring quantity of grassland vegetation. In: Measurement of Grassland Vegetation and Animal Production. Commonwealth Bureau of Pastures and Field Crops, Hurley, Berkshire, England. C . A . B . Bulletin #52. McCarthy, D . and Kearney, M . (1977). Evaluation of a simple method for measuring pasture dry matter production. Irish Grassland and Animal Production Association Journal. V o l . 12, p l29. Meijs, J . A . C (1981). Herbage intake by grazing dairy cows. Institute for Livestock Feeding and Nutrition Research. Tloorn" Lelyslad. Centre for Agricultural Publishing and Documentation, Wageningen 1981. Meijs, J . A . C ; Wallers, R.J .K. and Keen, A . (1982). Sward methods. In: Herbage Intake Handbook. Edited by J .D. Leaver. Published by the British Grassland Society, Hurley 16-36. Metson, A . J . ; Gibson, E .J . ; Hunt, J .L . and Saunders, W.M.II . (1979). Seasonal variations in chemical composition of pasture 111 Silicon, aluminum, iron, zinc, copper and manganese. N . Z . Journal of Agricultural Research. V o l . 22 309-318. 203 Melson, A J . ; Saunders, W . M . H . ; Collie, T . W . and Graham, V . W . (1966). Chemical compostion of pastures in relation to grass tetany in beef breeding cows. New Zealand J. Agr. Res. Vol.9, 410-436. Michalk, D . L . and Herbert, P .K. (1977). Assessment of four techniques for estimating yield on dryland pastures. Agronomy Journal. V o l . 69, 864-868 . Mitchell, R . L . (1964). Trace elements in soils. In: Chemistry of the Soil. (Ed) F . E . Bear. Pub. Reinhold, New York, U .S .A . 320-368. Morley, F . H . W . (1968). Pasture growth curves and grazing management. Australian Journal of Experimental Agriculture and Animal Husbandry. Vol .8 , 40-45. Morris, R . M . and Thomas, J . G . (1972). The seasonal pattern of dry matter production of grasses in the North Pennines. Journal of the British Grassland Society. Vol.28, 31-40. Morrison, J. (1980). The influence of climate and soil on the yield of grass and its response to fertilizer nitrogen. In: Prins, W . H . and Arnold, G . H (Eds.). The role of nitrogen in intensive grassland production. Pudoc, Wageningen, the Netherlands. 51-58. Morrison, J . ; Denehy, I L L . and Chapman, P.F. (1983). Possibilities for the strategic use of fertilizer N on white clover / grass swards. In: Corral 1, A . J . (Ed) Efficient Grassland Farming. Proceedings of the 9lh General Meeting of the European Grassland Federation published as Occasional Symposium of the British Grassland Society. No. 14, 227-231. Muhrer, M . E . ; Garner, G.B. ; Pfander, W . H . and O'Dell, B . L . (1956). The effect of nitrate on reproduction and lactation. Journal of Animal Science. V o l . 15, 1291-1292. Nevens, W . B . (1941). Cows urine as a fertilizer for Bluegrass Pastures. Journal of Dairy Science. Vol.24, 761-769. Nevens, W . B . (1943). The relation of the kind of pasture crop to the weed content of the forage. Journal of Dairy Science. Vol.26, 877-882. 204 Nevens, W . B . (1945). A comparison of sampling procedures in making pasture yield determinations. Journal of Dairy Science. Vol.28, 171-185. Nicholson, J . W . G . (1982). Mineral content of New Brunswick Forages. Mineral Nutrition Seminar, March 3, 1982, Moncton, N . B . 18pp. Nicol l , G . B . (1980). Evaluation of the machine washing of herbage samples cut at two lengths. Irish Journal of Agricultural Research. Vol.19, 91 -97 . Nicol l , G . B . (1982). Effect of level of soil contamination on machine-washed herbage samples. Irish Journal of Agricultural Research. Vol.21, 67-73. Norman, M . J . T . (1963). The pattern of dry matter and nutrient content changes in native pastures at Katherin, Northern Territories. Australian Journal of Agriculture and Animal Husbandry. Vol .3 , 119-124. Oohara, I-I.; Yoshida, N . ; Fukunaga, K. ; Colby, W . G . ; Drake, M . ; Wedin, W.F.; Kemmler, G . ; Uxekull, H.R. and Hasegawa, M . (1981). Impact of phosphorous and potassium fertilization on maintaining alfalfa-orchardgrass swards in Hokkaido, Japan. Proceedings of the X I V International Grasslands Congress (Lexington, U . S . A . ) 309-31 1. Orr, R.J . ; Parsons, A . J . ; Treacher, T . T . and Penning, P.D. (1988). Seasonal pattern of grass production under cutting or continuous stocking managements. Grass and Forage Science. Vol.43, 199-207. Oyenuga, V . A . (1960). Effect of stage of growth and frequency of cutting on the yield and chemical composition of some Nigerium fodder grasses Panicum maximum Jacq. Journal of Agricultural Science V o l . 55, 339-350. Parkinson, J . A . and Allen, S .E. , (1975). A wet oxidation procedure suitable for the determination of nitrogen and mineral nutrients in biological material. Commun. Soil Sci. Plant Anal. 6: pg 1-11. Perkin Elmer (1976). Analytical Methods for Atomic Absorption Spectophotometry. Pub. Perkin Elmer. 205 Pcrvival, N.S. and McClintock, M . B . (1982). Role of grazing management in manipulating the balance of ryegrass and paspalum in pastures. New Zealand Journal of Experimental Agriculture. V o l . 10, 365-370. Phillips, D . S . M . and Clark, S.E. (1971). The calibration of a weighted disc against pasture dry matter yield. Proceedings of the New Zealand Grassland Association. Vol.33, 68-75. Plummer, B . E . (1953). Chemical composition of grasses legumes in Maine. Maine Agricultural Experiment Station. University of Maine. Bulletin 513. 29pp. Powell, T . L . (1974). Evaluation of a weghted disc meter for pasture yield estimation on intensively slocked dairy pasture. New Zealand Journal of Experimental Agriculture. Vol .2 , 237-241. Priiis, W . H . (1978). Grassland productivity as affected by an intensive nitrogen and cutting regime in preceding years. A preliminary study. Proceedings of the 7th. General Meeting of the European Grassland Federation (Ghent). 10.23-10.27. Prins, W . H . ; Burg, P .F.J , van and Wieling, IT. (1980). The seasonal response of grassland to nitrogen at different intensities of nitrogen fertilization, with special reference to methods of response measurements. In: Prins, W . H . and Arnold, G . H . (Editors) The role of nitrogen in intensive grassland production. Pub. Pudoc. Wageningen, the Netherlands. Puffe, D . ; Morgner, F. and Zerr, W. (1984a). Untersuchungen zu den Gehalten an verschiedcnen Inhalstsstoffen wichtiger Futterpflanzen. 2. Mitteilung: Einfuhrung in die Versuchsfrage; Trockensubstanz-r ohprolein-und rohfascrgehalte Zeilschrift. V o l . 30, No. 1, 36-51. Puffe, D . ; Morgner, F. and Zerr, W . (1984b). Untersuchungen zu den Gehalten an verschiedenen Inhalstsstoffen wichtiger Futterpflanzen. 2. Mitteilung: Mineralsloffgehalle; kieselsauregehalt. Zeilschrift. V o l . 30, N o . l , 52-70. Radcliffe, J . E . (1971). Cutting techniques for pasture yields on hill country. Proceedings of the New Zealand Grassland Association. V o l . 33, 91-104. 206 Radcliffe, J . E . (1974). Seasonal distribution of pasture production in New Zealand. I. Methods of measurement. New Zealand Journal of Experimental Agriculture. V o l . 2, 337-340. Radcliffe, J . E . (1976). Seasonal distribution of pasture production in New Zealand. X . Rangitikei District. New Zealand Journal of Experimental Agriculture. V o l . 4, 163-170. Radcliffe, J . E . (1982). Effects of aspect and topography on pasture production in hill country. New Zealand Journal of Agricultural Research. V o l . 25, 485-496. Radford, P.J. (1967). Growth analysis formulae - their use and abuse. Crop Science. V o l . 7, 171-179. Reay, P.F. and Marsh, B . (1976). Element composition of ryegrass and red clover leaves during a growing season. N . Z . Journal of Agricultural Research. V o l . 19, 469-472. Reddy, G . D . ; Alston, A . M . ; Tiller, K . G . (1981). Seasonal changes in the concentrations of copper, molybdenum and sulfur in pasture plants. Aust. J-Exp. Agric. Anim. Husb. Vol 21, 498-505. Reid, D . (1959). Studies on the cutting management of grass clover swards. I. The effect of varying the closeness of cutting on the yields of an established grass-clover sward. Journal of Agricultural Science, Cambridge. V o l . 53, 299-312. Reid, D . (1966). The response of herbage yields and quality to a wide range of nitrogen application rates. Proceedings of the X International Grassland Congress (Helsinki) 209-213. Reid, D . (1985). A comparison of the yield responses of four grasses to a wide range of nitrogen application rales. Journal of Agricultural Science, Cambridge. V o l . 105, 381-387. Reid, D . and Castle, M . E . (1965). The response of grass-clover and pure grass leys to irrigation and fertilizer nitrogen treatment. I. Irrigation treatment. Journal of Agricultural Science, Cambridge. V o l . 64, 185-194. Reid, R . L . ; Post, A . J . ; Jung, C A . (1970). Mineral composition of forages. West Virginia Agricultural Experiment Station Bulletin 589 T. 35pp. 207 Reid.D. and Maclusky, D.S. (1960). Studies on the cutting management of grass clover swards. II. The effects of close cutting with either a gang mower or a reciprocating knife mower on the yields of an established grass-clover sward. Journal of Agricultural Science, Cambridge. V o l . 54, 158-165. Reith, J .W.S. and Mitchell, R . L . (1964). The effects of soil treatment on trace element uptake by plants. In: Plant Analysis and Fertilizer Problems IV. American Society of Horticultural Science, College Park, Maryland. 241-254. Richardson, A . E . V . (1932). Investigations on irrigated pastures. I. The yield and botanical composition of an irrigated permanent pasture under various systems of pasture management. C l . Sci. Ind. Res. Aust. Bulletin #71, 11-36. Robert, R . A . and Hunt, I.V. (1936). The effect of shoot cutting on the growth root and shoot of perennial ryegrass (Lolium perenne L . ) and of Timothy (Phleum pratensc L . ) . Welsh Journal of Agriculture. Vo l . 12, 158-174. (cited from Herbage Abstracts, 1936. V o l . 6, 375). Robinson, R.R. and Sprague, V . C . (1947). The clover populations and yields of a Kentucky bluegrass sod as affected by nitrogen fertilization, clipping treatments and irrigation. Agronomy Journal. V o l . 39, 107-116. Robinson, R.R.; Pierre, W . H . and Ackerman, R . A . (1937). A comparison of grazing and clipping for determining the response of permanent pastures to fertilizer. Journal of the American Society of Agronomy. V o l . 29, 349-359. Robinson, R.R.; Sprague, V . C . and Lueck, A . G . (1952). The effect of irrigation, nitrogen fertilization and clipping treatments on persistence of clover and on seasonal distribution of yields in a Kentucky bluegrass sod. Agronomy Journal. V o l . 44, 239. ryegrass / white clover sward. Grass and Forage Science, Vol.42, 429 -438 . Scars, P .D. (1944). Pasture plot measurement technique. A study of Ryegrass and Clover strains and phosphate response when meaured under different systems of plot measurement. New Zealand Journal of Science and Technology. Section A . Vol.25. 177-190. 208 Sears, P.D. and Goodall, V . C . (1942). Apparatus for the collection of urine and faeces from grazing sheep. New Zealand Journal of Science and Technology. Section A . Vol.23. 301-304. Sharrow, S .H . (1983). Forage standing crop and animal diets under rotational vs. continuous grazing. Journal of Range Management. V o l . 36, 447-450. Shaw, N.l-L; Elich, T . W . ; Haydock, K .P . and Waite, R.B. (1965). A comparison of 17 introductions of Paspalum species and naturalised Pas pa Ium dilatation under cutting at Samford, S. E . Queensland. Australian J Journal of Experimental Agriculture and Animal Husbandry. V o l . 5, 423 -432 . Shcehan, W. ; Fontenot, J.P. and Blascr, R . E . (1985). In-vitro dry matter digestibility and chemical composition of autumn-accumulated tall fescue, orchardgrass and red clover. Grass and Forage Science. V o l . 40. No.3. 317-322. Shelford, J . A . and Peterson, R . G . (1978) Relationships between feed mineral content, blood serum and dairy cow fertility. Proceedings of the 13th. Annual Pacific Northwest Animal Nutrition Conference (Vancouver). 108-118. Shrivastava, J.P.; Miller, D . A . and Jackobs, J .A . (1969). Estimating alfalfa yields. Agronomy Journal. V o l . 61, 649-651. Socst, P.J. van (1965). Symposium on factors influencing the voluntary intake of herbage by ruminants: voluntary intake in relation to chemical composition and digestibility. Journal of Animal Science. Vol .24 , 834-843. Stewart, A . B . and Holmes, W. (1953). Nitrogenous manuring of grassland. I. Some effects of heavy dressings of nitrogen on the mineral composition. J. Sci. Fd . Agric. Vol.4 401-408. Sullivan, J .T.; Phillips, T . G . ; Loughlin, M . E . and Sprague, V . G . (1956). Chemical composition of forage grasses. 2. Successive cutting during the growing season. Agronomy Journal. V o l . 48, 11.-14. • 20 9 Swift, R.W.; Cowan, R . L . ; Barron, G.P.; Maddy, K . H . ; Grose, E . C . and Washko, J .B . (1952). Further determinations of the nutritive value of forages. Journal of Animal Science. Vol.11. 389. Tallowin, J . R . B . and Brookman, S . K . E . (1988). Herbage potassium levels in a permanent pasture under grazing. Grass and Forage Science. Vol.43, 209-212 . Technicon Autoanalyzer II. (1974). Industrial Method No. 321-74A. Thomas, B.; Thompson, A . ; Oyenuga, V . A . and Armstrong, R . H . (1952). The ash constituents of some herbage plants at different stages of maturity. Empire Journal of Experimental Agriculture. V o l . 20. 10-22. Thompson, J .A . ; Sheridan, K . P . and Hamilton, B . A . (1976). The effects of rates of stocking with rotational grazing on the productivity of dryland lucerne at Tamworth, New South Wales. Australian Journal of Experimental Agriculture and Animal Husbandry. V o l . 16, 845-8 853. Todd, J.R. (1961). Magnesium in Forage plants. I. Magnesium content of different species and strains as affected by season and soil treatment. J. agric. Sci., Camb. V o l . 56. 411-415. Todd, J.R. (1966). Methods of increasing the magnesium contents of herbage with particular reference to the prevention of hypomagnesemic tetany in ruminants. In: Proceedings of the X International Grassland Congress (Helsinki, Finland) 178-180. Tucker, C . J . (1980). A critical reveiw of remote sensing and other methods for non-destructive estimation of standing crop biomass. Grass and Forage Science. V o l . 35, 177-182. Van Riper, G . E . and Smith, D. (1959). Changes in the chemical composlion of the herbage of alfalfa, medium red clover, ladino clover and bromegrass with advance in maturity. Agricultural Experiment Station. University of Wisconsin, Madison. Research Report No.4. 25pp . 210 Waddington, J. and Cooke, D . A . (1971). The influence of sample size on the precision and estimates of herbage production and consumption in two grazing experiments. Journal of the British Grassland Society. V o l . 26, 95-101. Waldern, D . E . (1971). A rapid micro-digestion procedure for neutral and acid detergent fibre. Can. J. Animal Science. 51 : 67. Wall , L . L . , Sr. and Gehrke, C . W . (1975). A n automated total protein nitrogen method. J. Assoc. Offic. Anal. Chem. 58 (6) : 1221-1226. Walters, R .J .K. and Evans, E . M . (1979). Evaluation of a sward sampling technique for estimating herbage intake by grazing sheep. Grass and Forage Science. V o l . 34, 37-44. Whitehead, D . C . (1966). Nutrient minerals in grassland herbage. Commonwealth Agricultural Bureaux. Mimeographed Publication No. 1 Commonwealth Bureau of Pastures and Field Crops, Hurley, Berkshire, U . K . 83pp. Whitehead, D . C . (1970). The role of nitrogen in grassland productivity. A review of information from temperate regions. Commonwealth Agricultural Bureaux. Commonwealth Bureau of Pastures and Field Crops, Hurley, Berkshire, England. Bulletin 48. 202pp. Williams, C . N . and Biddiscombe, E . F . (1964). Thr winter growth of selected pasture grasses. Australian Journal of Experimental Agriculture and Animal Husbandry. V o l . 4, 357-362. Wilm, H . G . ; Costello, D .F . and Klipple, G . E . (1944). Estimating forage yields by the double sampling method. Journal of the American Society of Agronomy. V o l . 36, 194-203. Wilman, D . (1965). The effect of nitrogenous fertilizer on the rate of growth of Italian ryegrass. Journal of the British Grassland Society. Vol .20 , 248-254. Wilman, D . and Wright, P.T. (1981). Some effects of applied nitrogen on grass growth in field swards at different times of year. Proceedings of the X I V International Grassland Congress (Lexington, Kentucky). 2 9 7 - 2 9 9 . Wit, C . E . de; Dijkshoorn, W.; and Noggle, J .C . (1963). Ionic balance and growth of plants. Versl. Landbouwk Onderz. No 69.15, 69. (cited from Whitehead, 1966). Wolfe, E . C . and Crofts, F . C . (1969). The effect of nitrogen fertilizer on the seasonal production of irrigated perennial grasses in coastal New South Wales. Australian Journal of Experimental Agriculture and Animal Husbandry. Vol .9 , 610-616. Wolton, K . M . (1963). A n investigation into the simulation of nutrient returns by the grazing animal in grassland experimentation. J. British Grassland Society. Vo l 18. 213-219. Woodman, H . E . and Norman, D . B . (1932). Nutritive value of pasture. IX. The influence of the intensity of grazing on the yield, composition and nutritive value of pasture herbage (part IV). V o l . 22, 852-873. Wright, M . J . and Davison, K . L . (1964). Nitrate accumulation in crops and nitrate poisoning in animals. Adv. Ag'ron. 197-248. Younie, D.; Heath, S.B.; Yackiminie, D.S. and Carr, D .W. (1986). Development of a low-input beef system based on while clover. In Frame, J. (ed.). Grazing. Occasional Symposium of the British Grassland Society. APPENDIX I Dry matter yields at each harvest Agassiz, 1984 - 1986 TABLE 47: YIELD FOR LOW-INFREQUENT CUTTING TREATMENT, AGASSIZ 1984 CULTIVAR HARVEST DATE TOTAL AND YEAR MAY 8 JUNE 5 JULY 3 AUG 21 SEPT 11 OCT 12 YIELD AUG 22 SEPT 21 Yield fka D.M./ha) 1984 PRAIRIAL 2,642 a 1,914 c 1,937 a 2,441 a 2,123 a 1,682 a 12,739 a SUMAS 1,908 b 1,841 c 1,895 a 2,239 a 1,892 a 1,545 a 11,320 b BARLANO 1,784 b 2,501 b 1,584 a 1,886 b 1,671 ab 197 b 9,623 c NORLEA 1,775 b 2,799 a 1,055 b 0 c 1,195 b 505 b 7,328 d /1EAN YIELD 2,027 2,264 1,618 1,641 1,720 982 10,252 ,E. OF MEAN 196 73 175 135 251 200 308 L.S.D. 443 166 395 306 568 453 697 Note: Different letters within the same column are. significantly different ( P = 0.05 ). TABLE 48: YIELD FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE TOTAL MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 YIELD Yield (kg D.M./ha) 1984 PRAIRIAL 1,610 a 891 b 1,512 a 2,584 a 1,799 a 1,761 a 1,411 a 356 a 11,924 SUMAS 1,295 b 1,041 b 1,417 a 2,844 a 1,737 b 1,267 b 1,166 a 180 b 10,947 BARLANO 996 c 1,416 a 1,321 a 1,694 b 1,640 b 1,115 be 1,128 a 0 c 9,310 NORLEA 829 c 1,506 a 1,564 a 1,494 b 1,160 b 748 c 436 b 0 c 7,737 MEAN YIELD 1,182 1,214 1,453 2,154 1,584 1,223 1,035 134 9,979 S.E. of the MEA 105 140 104 323 251 179 140 28 535 L.S.D. 237 317 234 730 569 405 316 63 1211 Note: Different letters within the same column are significantly different ( P = 0.05 ). TABLE 49: YIELD FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE TOTAL AND YEAR MAY 6 JUNE 4 JULY 3 AUG 5 SEPT 20 YIELD Yield (ka D.M./ha) 1985 PRAIRIAL 1,886 a 1,454 b 1,758 a 1,139 a 973 a 7,210 a SUMAS 1,443 ab 1,533 b 1,406 b 1,212 a 872 ab 6,466 a BARLANO 1,860 a 1,574 b 1,257 b 730 b 653 be 6,074 a NORLEA 767 b 1,744 a 1,229 b 0 c 446 c 4,186 b MEAN YIELD 1,489 1,576 1,412 770 736 5,984 S.E. OF MEAN 366 54 90 148 130 562 L.S.D. 829 123 203 334 295 1,272 Note: Different letters within the same column are significantly different ( P = 0.05 ). TABLE 50: YIELD FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE TOTAL MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 YIELD Yield (ka D.M./ha) 1985 PRAIRIAL 1,240 b 1,024 b 1,239 a 1,077 a 690 a 415 a 594 a 578 a 6,857 SUMAS 1,450 a 1,048 b 993 b 1,074 a 794 a 370 a 410 b 167 b 6,306 BARLANO 1,008 c 1,230 a 1,362 a 939 a 327 b 0 b 569 a 626 a 6,061 NORLEA 300 d 1,072 b 1,003 a 765 a 0 b 286 b 224 c 220 b 4,170 MEAN YIELD 999 1,093 1,224 964 . 453 268 449 398 5,848 S.E. of the MEAN 91 62 68 149 158 145 67 68 446 L.S.D. 207 140 153 337 357 328 151 154 1009 Note: Different letters within the same column are significantly different ( P = 0.05 ). T A B L E 51: YIELD FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1986 CULTIVAR HARVEST DATE TOTAL AND YEAR MAY 7 JUNE 7 JULY 8 AUG 8 SEPT 19 OCT 25 YIELD 1986 PRAIRIAL 1,401 a 3,152 c SUMAS 945 b 3,717 a BARLANO 781 c 3,694 a NORLEA 514 d 3,351 b Yield (kg D.M./ha) 2,523 a 1,457 a 1,864 2,306 a 1,038 c 1,492 2,720 a 1,032 c 745 2,354 a 1,200 b 552 a 1,589 a 11,986 a b 1,327 b 10,825 b c 677 c 9,649 c c 516 c 8,487 d MEAN YIELD 910 3,478 2,476 1,182 1,163 1,027 10,237 S.E. OF MEAN 71 75 144 60 143 77 199 L.S.D. 161 170 325 136 323 174 450 Note: Different letters within the same column are significantly different ( P = 0.05 ). to 218 APPENDIX TT Forage Quality information ai each harvest Agassiz, 1984 and 1985 TABLE 52 A: CRUDE PROTEIN CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT AGASSIZ. 1984 CULTIVAR HARVEST DATE YEARLY AND YEAR MAY 8 JUNE 5 JULY 3 AUG 2/ SEPT 7/ OCT 12 MEAN AUG 22 SEPT 21 PERCENT PROTEIN 1984 PRAIRIAL 18.49 26.85 18.52 24.41 23.38 23.27 22.49 SUMAS 20.05 24.60 19.97 22.60 22.94 23.44 22.27 BARLANO 17.33 22.45 16.61 21.22 24.09 25.21 21.15 NORLEA 17.43 20.56 16.59 - 23.55 26.19 20.86 RVEST MEAN 18.33 23.62 17.92 22.74 23.49 24.53 TABLE 52 B: CRUDE PROTEIN CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT ' AGASSIZ. 1984 CULTIVAR HARVEST DATE CULTIVAR MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 MEAN PERCENT PROTEIN 1984 -PRAIRIAL 19.25 26.91 25.15 14.78 26.25 24.74 25.16 25.40 23.46 SUMAS 20.73 25.83 25.68 16.68 23.85 23.82 25.90. 25.23 23.47 BARLANO 18.46 25.04 21.96 11.34 23.42 25.65 24.71 - 21.51 NORLEA 19.24 26.27 22.58 14.14 23.61 26.94 27.03 - 22.83 HARVEST MEAN 19.42 26.01 23.84 14.24 24.28 25.29 25.70 25.32 TABLE 52 C: CRUDE PROTEIN CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT AGASSIZ. 1985 CULTIVAR HARVEST DATE YEARLY AND YEAR MAY 6 JUNE 4 JULY 3 AUG 5 SEPT 20 MEAN PERCENT PROTEIN 1985 PRAIRIAL 27.80 21.78 21.59 17.13 23.82 22.42 SUMAS 21.74 18.88 22.66 17.50 23.40 20.84 BARLANO 26.58 18.88 22.44 19.16 24.54 22.32 NORLEA 26.19 22.36 24.63 16.00 22.30 HARVEST MEAN 25.58 20.48 22.83 17.93 21.94 TABLE 52 D: CRUDE PROTEIN CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 MEAN PERCENT PROTEIN 1985 PRAIRIAL 27.11 24.67 18.76 24.04 20.17 18.66 23.33 27.77 23.06 SUMAS 24.01 23.25 21.50 25.85 20.10 20.22 23.76 25.37 23.01 BARLANO 26.30 22.53 17.88 25.22 20.18 - 22.39 25.22 22.82 NORLEA 26.16 26.26 21.82 25.96 18.79 26.72 25.33 24.43 HARVEST MEAN 25.90 24.18 19.99 25.27 20.15 19.22 24.05 25.92 TABLE 53 A: CRUDE PROTEIN YIELD FOR LOW-INFREQUENT CUTTING TREATMENT AGASSIZ 1984 CULTIVAR AND YEAR MAY 8 JUNE 5 HARVEST JULY 3 DATE AUG 2/ AUG 22 SEPT 11 SEPT 21 OCT 12 TOTAL YIELD — _ Yield (ka D.M./ha) 1984 PRAIRIAL 489 514 359 596 496 391 2,845 SUMAS 383 453 378 506 434 362 2,516 BARLANO 309 561 263 400 403 50 1,986 NORLEA 309 575 175 0 281 132 1,474 HARVEST MEAN 372 526 294 376 404 234 TABLE 53 B: CRUDE PROTEIN YIELD FOR HIGH-FREQUENT CUTTING TREATMENT, AGASSIZ 1984 CULTIVAR HARVEST DATE TOTAL MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 YIELD PROTEIN YIELD (kg D.M. / ha) 1984 PRAIRIAL 310 240 380 382 472 436 355 90 2665 SUMAS 268 269 364 474 414 302 302 45 2439 BARLANO 184 355 290 192 384 286 279 0 1969 NORLEA 159 396 353 211 274 202 118 0 1713 HARVEST MEAN 230 315 347 315 386 306 263 34 2197 TABLE 53 C: CRUDE PROTEIN YIELD FOR LOW-INFREQUENT CUTTING TREATMENT AGASSIZ. 1985 CULTIVAR HARVEST DATE TOTAL AND YEAR MAY 6 JUNE 4 JULY 3 AUG 5 SEPT 20 YIELD Yield (kg D.M./ha) 1985 PRAIRIAL 524 . 317 380 195 232 - 1,647 SUMAS 314 289 319 212 204 - 1,338 BARLANO 494 297 282 140 160 - 1,374 NORLEA 201 390 303 0 71 - 965 HARVEST MEAN 383 323 321 137 167 T A B L E 53 D: CRUDE PROTEIN YIELD FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE TOTAL MAY 6 MAY 27 JUNE 18 . JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 YIELD PROTEIN YIELD fka P.M. / ha) 1985 PRAIRIAL 336 253 232 259 139 77 139 161 1596 SUMAS 348 244 213 278 160 75 97 42 1457 BARLANO 265 277 244 237 66 0 127 158 1374 NORLEA 78 282 219 199 0 54 60 56 947 VEST MEAN 257 264 227 243 91 51 106 104 1343 is) to CO TABLE 54 A: A.D.F. CONTENT FOR LOW - INFREQUENT CUTTING TREATMENT, AGASSIZ 1984 CULTIVAR HARVEST DATE YEARLY AND YEAR MAY 8 JUNE 5 JULY 3 AUG 2/ SEPT 7/ OCT 12 MEAN AUG 22 SEPT 21 PERCENT ADF 1984 PRAIRIAL 31.22 29.87 35.68 31.65 34.90 30.80 32.35 SUMAS 27.67 29.34 34.08 31.99 33.64 29.51 31.04 BARLANO 25.30 30.88 32.65 33.14 28.22 22.04 28.71 NORLEA 23.68 30.02 28.25 - 29.94 25.04 27.39 RVEST MEAN 26.97 30.03 32.67 32.26 31.68 26.85 TABLE 54 B A.D.F. CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR MAY 7 MAY 28 HARVEST DATE JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 CULTIVAR MEAN PERCENT A.D.F. 1984 PRAIRIAL 25.60 29.04 31.07 40.21 30.72 31.52 29.16 28.78 30.76 SUMAS 24.18 29.34 28.10 39.69 32.24 31.50 25.48 28.54 29.88 BARLANO 21.69 28.67 28.95 36.20 29.81 28.25 29.86 - 29.06 NORLEA 21.22 27.09 28.84 33.42 28.98 26.44 26.49 - 27.50 ,RVEST MEAN 23.17 28.54 29.24 37.38 30.44 29.43 27.75 28.66 to to oo TABLE 54 C : A.D.F. CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE AVERAGE AND YEAR MAY 6 JUNE 4 JULY 3 AUG 5 SEPT 20 YIELD PERCENT A.D.F. 1985 PRAIRIAL 22.41 35.93 30.02 36.05 33.40 31.56 SUMAS 31.88 36.39 28.75 32.99 33.16 32.63 BARLANO 23.32 34.13 24.77 29.87 25.09 27.44 NORLEA 19.88 32.01 24.81 31.43 27.03 MEAN YIELD 24.37 34.62 27.09 32.97 30.77 TABLE 54 D: A.D.F. CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 MEAN PERCENT A.D.F. 1985 PRAIRIAL 23.01 32.59 32.23 29.97 35.49 34.77 27.53 25.25 30.11 SUMAS 23.11 32.23 30.49 29.11 35.13 31.80 26.09 23.13 28.89 BARLANO 21.49 32.19 28.69 25.80 29.63 - 24.40 23.91 26.59 NORLEA 17.39 27.92 26.43 24.22 - 30.47 22.04 20.73 24.17 RVEST MEAN 21.25 31.23 29.46 27.28 33.42 32.35 25.02 23.26 TABLE 55A: ASH CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE AND YEAR MAY 8 JUNE 5 JULY 3 AUG 21 SEPT 7/ AUG 22 SEPT 21 OCT 12 YEARLY MEAN 1984 PRAIRIAL SUMAS BARLANO NORLEA 9.44 8.38 9.01 9.36 PERCENT ASH 9.22 9.64 9.98 9.64 9.26 9.75 9.69 9.09 9.07 9.69 9.34 9.31 9.99 9.16 9.29 8.48 8.35 8.64 9.49 9.19 9.34 9.19 9.38 HARVEST MEAN 9.05 9.54 9.53 9.37 9.44 8.74 TABLE 55B: ASH CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE CULTIVAR MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 MEAN PERCENT ASH 1984 PRAIRIAL 8.51 9.40 9.55 8.78 7.87 8.78 9.20 8.42 8.81 SUMAS 7.83 9.90 9.42 9.36 8.51 8.80 9.16 8.37 8.92 BARLANO 8.52 9.54 9.35 9.56 8.42 9.20 9.54 - 9.16 NORLEA 9.27 9.33 9.61 9.11 8.56 9.70 9.15 - 9.25 HARVEST MEAN 8.53 9.54 9.48 9.20 8.34 9.12 9.26 8.40 TABLE 55C: ASH CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY AND YEAR MAY 6 JUNE 4 JULY 3 AUG 5 SEPT 20 MEAN PERCENT ASH 1985 PRAIRIAL 8.95 9.06 8.56 6.93 7.45 8.19 SUMAS 8.85 8.82 8.76 7.17 9.03 8.53 BARLANO 9.33 9.47 9.06 9.04 8.99 9.18 NORLEA 9.36 9.39 9.38 7.09 8.81 HARVEST MEAN 9.12 9.19 8.94 7.71 8.14 TABLE 55D: ASH CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 MEAN PERCENT ASH 1985 PRAIRIAL 8.80 9.37 9.56 7.39 7.38 6.81 7.63 9.76 8.34 SUMAS 8.28 9.52 9.50 7.67 6.76 6.95 7.97 . 9.43 8.26 BARLANO 9.14 9.95 9.48 9.75 8.79 - 8.71 9.31 9.30 NORLEA 8.74 9.20 9.99 9.47 8.06 8.90 9.07 9.06 HARVEST MEAN 8.74 9.51 9.63 8.57 7.64 7.27 8.30 9.39 TABLE 56A: CALCIUM CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE AND YEAR MAY 8 JUNE 5 JULY 3 AUG 21 SEPT ll AUG 22 SEPT 21 OCT 12 YEARLY MEAN 1984 PRAIRIAL SUMAS BARLANO NORLEA 0.403 0.423 0.612 0.630 0.471 0.425 0.549 0.533 PERCENT CALCIUM 0.558 0.538 0.662 0.636 0.438 0.506 0.698 0.579 0.618 0.626 0.755. 0.662 0.591 0.569 0.574 0.519 0.517 0.619 0.626 HARVEST MEAN 0.517 0.494 0.599 0.547 0.645 0.599 TABLE 56B: CALCIUM CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE CULTIVAR MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 MEAN PERCENT CALCIUM 1984 PRAIRIAL 0.363 0.361 0.425 0.537 0.431 0.507 0.601 0.552 0.472 SUMAS 0.331 0.315 0.372 0.630 0.645 0.531 0.641 0.583 0.506 BARLANO 0.592 0.534 0.582 0.724 0.744 0.480 0.657 - 0.616 NORLEA 0.686 0.581 0.649 0.773 0.910 0.466 0.619 - 0.669 HARVEST MEAN 0.493 0.448 0.507 0.666 0.683 0.496 0.629 0.568 TABLE 56C: CALCIUM CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY AND YEAR MAY 6 JUNE 4 JULY 3 AUG 5 SEPT 20 MEAN PERCENT CALCIUM 1985 PRAIRIAL 0.514 0.289 0.504 0.579 0.800 0.537 SUMAS 0.467 0.268 0.494 0.583 0.936 0.550 BARLANO 0.708 0.520 0.637 0.854 0.737 0.691 NORLEA 0.572 0.469 0.652 - 0.879 0.643 HARVEST MEAN C.565 0.387 0.572 0.672 0.838 -TABLE 56D: CALCIUM CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 MEAN PERCENT CALCIUM 1985 PRAIRIAL 0.421 0.290 0.381 0.393 0.772 0.701 0.534 0.550 0.505 SUMAS 0.351 0.268 0.469 0.464 0.826 0.896 0.682 0.550 0.563 BARLANO 0.627 0.522 0.707 0.647 1.264 - 0.612 0.485 0.695 NORLEA 0.524 0.445 0.589 0.585 1.102 0.471 0.458 0.596 HARVEST MEAN 0.481 0.381 0.537 0.522 0.954 0.900 0.575 0.511 to to TABLE 57A: PHOSPHORUS CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE AND YEAR MAY 8 JUNE 5 JULY 3 AUG 21 SEPT 7/ AUG 22 SEPT 21 OCT 12 YEARLY MEAN 1984 PRAIRIAL SUMAS BARLANO NORLEA 0.316 0.290 0.326 0.323 0.352 0.349 0.376 0.367 PERCENT PHOSPHORUS 0.342 0.329 0.351 0.362 0.357 0.355 0.388 0.411 0.407 0.385 0.391 0.473 0.408 0.353 0.366 0.375 0.356 0.363 0.362 HARVEST MEAN 0.314 0.361 0.346 0.367 0.399 0.400 TABLE 57B: PHOSPHORUS CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE CULTIVAR MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 MEAN PERCENT PHOSPHORUS 1984 PRAIRIAL 0.304 0.376 0.380 0.297 0.338 0.407 0.355 0.324 0.348 SUMAS 0.291 0.352 0.343 0.302 0.330 0.360 0.316 0.302 0.325 BARLANO 0.309 0.384 0.440 0.300 0.344 0.427 0.434 - 0.377 NORLEA 0.317 0.408 0.469 0.370 0.352 0.440 0.406 0.395 HARVEST MEAN 0.305 0.380 0.408 0.317 0.341 0.409 0.378 0.313 TABLE 57C: PHOSPHORUS CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR AND YEAR MAY 6 JUNE 4 HARVEST DATE JULY 3 AUG 5 SEPT 20 YEARLY MEAN PERCENT PHOSPHORUS 1985 PRAIRIAL 0.370 0.410 0.361 0.284 0.330 0.351 SUMAS 0.291 0.350 0.351 0.322 0.337 0.330 BARLANO 0.409 0.434 0.374 0.310 0.359 0.377 NORLEA 0.409 0.491 0.366 0.219 0.371 HARVEST MEAN 0.370 0.421 0.363 0.305 0.311 TABLE 57D: PHOSPHORUS CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 MEAN PERCENT PHOSPHORUS 1985 PRAIRIAL 0.400 0.358 0.362 0.344 0.358 0.239 0.386 0.372 0.352 SUMAS 0.363 0.321 0.342 0.324 0.333 0.268 0.361 . 0.355 0.333 BARLANO 0.432 0.396 0.383 0.350 0.269 - 0.377 0.362 0.367 NORLEA 0.451 0.343 0.387 0.339 - 0.224 0.352 0.334 0.347 HARVEST MEAN 0.412 0.355 0.369 0.339 0.320 0.244 0.369 0.356 TABLE 58A: POTASSIUM CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE YEARLY AND YEAR MAY 8 JUNE 5 JULY 3 AUG 2/ SEPT 71 OCT 12 MEAN AUG 22 SEPT 21 PERCENT POTASSIUM 1984 PRAIRIAL 4.01 4.90 3.64 3.76 3.47 3.59 3.90 SUMAS 3.99 4.94 3.78 3.75 3.82 3.63 3.98 BARLANO 3.65 4.22 3.54 3.65 3.23 3.10 3.57 NORLEA 4.32 4.33 3.63 - 3.34 3.13 3.75 HARVEST MEAN 3.99 4.60 3.65 3.72 3.47 3.36 TABLE 58B: POTASSIUM CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE CULTIVAR MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 MEAN PERCENT POTASSIUM 1984 PRAIRIAL 3.57 4.42 3.94 3.18 3.00 3.61 3.24 3.15 3.51 SUMAS 3.33 4.26 3.91 3.39 3.31 3.57 3.02 2.97 3.47 BARLANO 3.52 3.92 3.99 3.07 3.12 3.13 3.25 - 3.43 NORLEA 3.86 4.61 4.12 3.85 3.65 3.22 3.15 - 3.78 HARVEST MEAN 3.57 4.30 3.99 3.37 3.27 3.38 3.17 3.06 TABLE 58C: POTASSIUM CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR AND YEAR MAY 6 JUNE 4 HARVEST DATE JULY 3 AUG 5 SEPT 20 YEARLY MEAN PERCENT POTASSIUM 1985 PRAIRIAL 2.81 3.83 2.69 1.53 1.55 2.48 SUMAS 2.86 3.76 3.30 1.88 1.62 2.68 BARLANO 3.17 4.08 3.02 1.80 2.31 2.88 NORLEA 3.45 4.65 3.11 1.37 3.15 HARVEST MEAN 3.07 4.08 3.03 1.74 1.71 TABLE 58D: POTASSIUM CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 MEAN PERCENT POTASSIUM 1985 PRAIRIAL 3.22 3.75 3.87 2.37 1.46 2.03 2.58 3:04 2.79 SUMAS 3.07 3.44 4.07 2.75 1.99 2.07 3.18" 2.89 2.93 BARLANO 3.55 4.24 3.84 2.83 1.66 - 2.52 3.23 3.12 NORLEA 3.18 4.96 4.83 3.10 1.74 3.67 3.90 3.63 HARVEST MEAN 3.26 4.10 4.15 2.76 1.70 1.95 2.99 3.27 TABLE 59A: SODIUM CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR AND YEAR MAY 8 JUNE 5 HARVEST DATE JULY 3 AUG 21 AUG 22 SEPT 11 SEPT 21 OCT 12 YEARLY MEAN PERCENT SODIUM 1984 PRAIRIAL 0.198 0.254 0.241 0.154 0.289 0.319 0.243 SUMAS 0.164 0.122 0.124 0.057 0.092 0.156 0.119 BARLANO 0.186 0.172 0.154 0.186 0.274 0.243 0.203 NORLEA 0.093 0.063 0.060 0.054 0.084 0.071 HARVEST MEAN 0.160 0.153 0.145 0.132 0.177 0.201 TABLE 59B: SODIUM CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE CULTIVAR MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 MEAN PERCENT SODIUM 1984 PRAIRIAL 0.179 0.180 0.292 0.209 0.170 0.239 0.319 0.289 0.235 SUMAS 0.166 0.133 0.154 0.124 0.060 0.105 0.169 0.152 0.133 BARLANO 0.141 0.137 0.139 0.113 0.154 0.187 0.233 - 0.158 NORLEA 0.068 0.085 0.069 0.058 0.061 0.053 0.081 - 0.068 HARVEST MEAN 0.139 0.134 0.164 0.126 0.111 0.146 0.201 0.221 TABLE 59C: SODIUM CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY AND YEAR MAY 6 JUNE 4 JULY 3 AUG 5 SEPT 20 MEAN PERCENT SODIUM 1985 PRAIRIAL 0.179 0.224 0.189 0.174 0.230 0.199 SUMAS 0.129 0.081 0.079 0.050 0.070 0.082 BARLANO 0.194 0.167 0.152 0.176 0.229 0.184 NORLEA 0.056 0.056 0.057 - 0.065 0.059 HARVEST MEAN 0.140 0.132 0.119 0.133 0.149 TABLE 59D: SODIUM CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 MEAN PERCENT SODIUM 1985 PRAIRIAL 0.181 0.178 0.231 0.164 0.152 0.198 0.232 0.234 0.196 SUMAS 0.122 0.063 0.107 0.071 0.064 0.082 0.083 0.103 0.087 BARLANO 0.156 0.153 0.203 0.172 0.193 - 0.224 0.275 0.197 NORLEA 0.044 0.048 0.040 0.037 - 0.064 0.049 0.061 0.049 HARVEST MEAN 0.126 0.111 0.145 0.111 0.136 0.115 0.147 0.168 TABLE 60A: MAGNESIUM CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE AND YEAR MAY 8 JUNE 6 JULY 2 AUG 2 / SEPT 7 / AUG 22 SEPT 21 OCT 12 YEARLY MEAN 1984 PRAIRIAL SUMAS BARLANO NORLEA 0.164 0.187 0.198 0.196 0.248 0.264 0.216 0.189 PERCENT MAGNESIUM 0.237 0.266 0.293 0.242 0.245 0.261 0.288 0.260 0.282 0.304 0.263 0.292 0.292 0.239 0.231 0.241 0.259 0.256 0.224 HARVEST MEAN 0.186 0.229 0.260 0.265 0.277 0.264 TABLE 60B: MAGNESIUM CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE CULTIVAR MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 MEAN PERCENT MAGNESIUM 1984 PRAIRIAL 0.137 0.212 0.218 0.205 0.211 0.237 0.266 0.257 0.218 SUMAS 0.155 0.202 0.225 0.283 0.261 0.235 0.314 . 0.279 0.244 BARLANO 0.178 0.196 0.201 0.207 0.198 0.279 0.263 - 0.217 NORLEA 0.160 0.180 0.211 0.223 0.189 0.187 0.205 - 0.194 HARVEST MEAN 0.158 0.198 0.214 0.230 0.215 0.235 0.262 0.268 TABLE 60C: MAGNESIUM CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY AND YEAR MAY 6 JUNE 4 JULY 3 AUG 5 SEPT 20 MEAN PERCENT MAGNESIUM 1985 PRAIRIAL 0.177 0.180 0.231 0.276 0.241 - 0.221 SUMAS 0.190 0.187 0.262 0.303 0.320 - 0.252 BARLANO 0.210 0.216 0.298 0.383 0.276 - 0.277 NORLEA 0.171 0.183 0.264 - 0.251 - 0.217 HARVEST MEAN 0.187 0.192 0.264 0.321 0.272 TABLE 60D: MAGNESIUM CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE CULTIVAR MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 MEAN PERCENT MAGNESIUM 1985 PRAIRIAL 0.159 0.179 0.207 0.195 0.252 0.215 0.202 0.292 0.213 SUMAS 0.152 0.183 0.255 0.267 0.306 0.303 0.235 - 0.290 0.249 BARLANO 0.197 0.211 0.264 0.279 0.386 - 0.264 0.278 0.258 NORLEA 0.162 0.159 0.220 0.216 0.242 0.183 0.207 0.198 HARVEST MEAN 0.168 0.183 0.237 0.239 0.315 0.253 0.221 0.267 TABLE 61A: COPPER CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE YEARLY AND YEAR MAY 8 JUNE 6 JULY 2 AUG 2 / SEPT 7 / OCT 12 MEAN AUG 22 SEPT 21 COPPER CONTENT (ma/ka) 1984 PRAIRIAL 12.0 16.4 12.7 15.3 22.5 15.3 15.7 SUMAS 11.0 16.6 13.2 16.5 14.2 15.8 14.6 BARLANO 12.3 14.4 9.7 17.2 12.8 12.1 13.1 NORLEA 9.0 13.3 8.7 - 15.0 13.1 11.8 HARVEST MEAN 11.1 15.2 11.1 16.3 16.1 14.1 TABLE 61B: COPPER CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE CULTIVAR MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 MEAN COPPER CONTENT ( mg/kg) 1984 PRAIRIAL 13.4 15.3 21.8 12.0 23.7 19.7 14.7 12.4 16.6 SUMAS 11.2 13.3 17.8 13.9 22.1 15.3 " -13.3 11.9 14.9 BARLANO 11.1 12.2 14.1 10.8 10.7 10.9 10.7 - 11.5 NORLEA 12.3 12.2 17.5 8.8 9.9 13.0 9.4 - 11.9 HARVEST MEAN 12.0 13.3 17.8 11.4 16.6 14.7 12.0 12.2 TABLE 61C: COPPER CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY AND YEAR MAY 6 JUNE 4 JULY 3 AUG 5 SEPT 20 MEAN COPPER CONTENT (ma/kcn 1985 PRAIRIAL 11.9 11.1 22.1 13.5 16.9 15.1 SUMAS 11.1 10.9 15.4 15.0 21.6 14.8 BARLANO 11.9 11.1 13.8 14.6 14.7 13.2 NORLEA 11.9 10.6 15.9 21.3 14.9 HARVEST MEAN 11.7 10.9 16.8 14.4 18.6 TABLE 61D: COPPER CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE CULTIVAR MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 MEAN COPPER CONTENT fma/ka) 1985 PRAIRIAL 21.3 17.5 12.9 13.5 14.1 21.4 15.6 18.8 16.9 SUMAS 21.9 12.6 12.2 11.3 13.4 24.2 .18.0 19.8 16.7 BARLANO 19.2 13.3 12.5 10.2 17.8 - 14.9~ 12.8 15.4 NORLEA 18.6 13.2 12.5 12.0 - 17.4 17.3 19.6 15.8 RVEST MEAN 20.3 14.2 12.5 11.8 15.1 21.0 16.5 17.8 to cc TABLE 62A: MANGANESE CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE AND YEAR MAY 8 JUNE 6 JULY 2 AUG 2 / SEPT 7 / AUG 22 SEPT 21 OCT 12 YEARLY MEAN 1984 PRAIRIAL SUMAS BARLANO NORLEA 102 68 66 73 MANGANESE CONTENT (mg/kg) 115 84 78 81 119 85 133 124 120 96 132 165 132 134 130 149 100 135 119 128 94 113 106 HARVEST MEAN 78 89 115 116 140 126 TABLE 62B: MANGANESE CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1984 CULTIVAR HARVEST DATE CULTIVAR MAY 7 MAY 28 JUNE 18 JULY 19 AUG 10 SEPT 4 OCT 2 OCT 29 MEAN MANGANESE CONTENT (mg/kg) 1984 PRAIRIAL 85 113 108 105 116 142 133 111 114 SUMAS 62 71 84 88 97 117 99 88 88 BARLANO 58 52 94 108 132 150 140 - 105 NORLEA 73 69 105 108 125 130 123 - 104 HARVEST MEAN 69 76 98 102 118 135 124 100 TABLE 62C: MANGANESE CONTENT FOR LOW-INFREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE YEARLY AND YEAR MAY 6 JUNE 4 JULY 3 AUG 5 SEPT 20 MEAN MANGANESE CONTENT (ma/ken 1985 PRAIRIAL 100 119 123 126 157 - 125 SUMAS 90 93 111 117 163 - 115 BARLANO 110 106 130 163 175 - 137 NORLEA 109 110 141 - 134 - 124 HARVEST MEAN 102 107 126 135 157 TABLE 62D: MANGANESE CONTENT FOR HIGH-FREQUENT CUTTING TREATMENT. AGASSIZ 1985 CULTIVAR HARVEST DATE CULTIVAR MAY 6 MAY 27 JUNE 18 JULY 10 AUG 5 AUG 30 SEPT 24 OCT 25 MEAN MANGANESE CONTENT (ma/ka) 1985 PRAIRIAL 92 121 119 118 117 111 163 193 129 SUMAS 66 94 92 104 116 115 157 184 116 BARLANO 101 103 112 127 164 - 150 186 135 NORLEA 115 104 111 119 - 135 127 157 124 HARVEST MEAN 93 106 108 117 132 120 149 180