Published online 5 September 2006
Published in Agron J 98:1333-1339 (2006)
DOI: 10.2134/agronj2005.0236
© 2006 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
Fertilizer Management
Irrigated Annual Ryegrass Responses to Nitrogen and Phosphorus on Calcareous Soil
Hagen Lippkea,*,
Vincent A. Habyb and
Tony L. Provinc
a Texas Agric. Exp. Stn., 1619 Garner Field Rd., Uvalde, TX 78801
b Texas Agric. Exp. Stn., P.O. Box 200, Overton, TX 75684
c Dep. of Soil and Crop Science, Texas A&M Univ., College Station, TX 77843
* Corresponding author (h-lippke{at}tamu.edu)
Received for publication August 17, 2005.
 |
ABSTRACT
|
|---|
Annual ryegrass (Lolium multiflorum Lam.) is increasingly being planted under irrigation as forage for stocker cattle in south central USA, but fertilizer recommendations for maximum production are not well defined. Labile N and P are generally deficient in this region's calcareous soils; K availability is usually considered adequate for plant growth. Growth responses of annual ryegrass to a factorial array of five levels of N and four levels of P were studied for 3 yr on a Knippa clay soil (fine, mixed, superactive, thermic Vertic Calciustolls). Responses to N and P were generally curvilinear, but they also had a strong linear component in Year 1. Increments of growth responses to levels of fertilizer above 269 kg applied N or 39 kg applied P ha–1 were relatively small. Response surface models projected maximum annual dry matter yields that averaged 9.23 t ha–1 and were associated with predicted requirements for applied N and P that averaged 488 and 61 kg ha–1, respectively. Economically optimal levels of applied N were predicted to range from 250 to 315 kg ha–1; the predicted economically optimal range for applied P was 31 to 41 kg ha–1. Nitrogen concentration in ryegrass forage increased as applied N increased. The data suggest that fertilizing vegetative ryegrass to maintain N in leaf tissue 
32 g kg–1 provides economically optimal growth for both the ryegrass crop and the young cattle grazing it.
Abbreviations: CP, crude protein • DM, dry matter • NA, applied nitrogen • NS, initial soil nitrate nitrogen • PA, applied phosphorus • PS, initial available soil phosphorus • Yr, year
|
INTRODUCTION
|
|---|
ANNUAL Italian ryegrass is becoming an important species for grazing stocker cattle in the south central USA, particularly on irrigated land. Data for yield responses to applied fertilizer for annual ryegrass grown under irrigation on alkaline soils of this region are not available. Fertilizer recommendations from soil testing laboratories for irrigated ryegrass production in this region have been based on research data from nonirrigated ryegrass fertilizer trials at other locations (Robinson et al., 1987; Westfall et al., 1971).
Annual ryegrass yields in southeastern USA have generally increased with total annual N application rates as high as 450 kg ha–1, which is usually the highest rate applied (Allen et al., 1974; Dunivan, 1975; Robinson et al., 1987). Hovermale (1993) obtained maximum ryegrass yields with 380 kg N ha–1; however, the maximum yield was only 6.7 t ha–1. Robinson et al. (1987) compared N rates up to 560 kg ha–1 by 112-kg ha–1 increments applied as ammonium nitrate. Significant yield increases at N rates to 336 kg ha–1 were observed in 1 yr and to 448 kg ha–1 in 2 yr. Three-year mean yields were higher at 448 kg N ha–1 than at 336 kg N ha–1 and 90% of the maximum yield, or 10.7 t ha–1, occurred at an average of 390 kg N ha–1 during the 3 yr.
The potential effect of fertilizer N on forage CP level is also a concern, because CP is inherently a major factor in growth rate of young grazing ruminants. The CP concentration in ryegrass needed for high performance is increased when grazed, because efficiency of utilization of CP is particularly low in fresh forages (Beever et al., 1974). Stocker cattle have demonstrated growth rates >1.2 kg d–1 (Lippke et al., 2000; Worrell et al., 1990). Calculations from FORAGVAL (Lippke and Herd, 1990) suggest that CP in grazed forage must be at least 20% (32 g N kg–1 forage) to support high levels of growth.
The majority of data on the response of annual ryegrass to P has been derived from research on acid soils, usually in conjunction with limestone treatments applied to elevate soil pH to the slightly acidic levels that are more suitable for increased grass production. Haby and Robinson (1997) summarized currently available research on ryegrass response to residual and applied P, noting that response to applied P on P-deficient soils is highly dependent on soil chemical properties. Hillard et al. (1992) reported a gradual increase in ryegrass forage yield from 3.2 to 5.1 t ha–1 as the rate of applied P was increased from zero to 480 kg ha–1 in two split applications to a Lilbert loamy fine sand (Arenic Plinthic Paleudult) with an average pH of 4.6 and 1 M KCl exchangeable Al concentration averaging 15 mg kg–1. In the same season, ryegrass yield was increased to an average of 8.4 t ha–1 where 3.8 t of limestone (64% effective CaCO3 equivalence) ha–1 had been applied and incorporated 2 yr earlier, raising pH to 6.3 and lowering exchangeable Al to less than 1 mg kg–1. Ryegrass response to applied P ranged from 8.3 to 8.6 t ha–1 as extractable soil P (Hons et al., 1990) increased from 3 to 55 mg kg–1 (Hillard et al., 1992). Robinson and Eilers (1996) obtained 90% of maximum ryegrass DM yield, or 9.8 t ha–1, with 45 to 50 kg P ha–1 on a Tangi silt loam (Typic Fragiudult) at pH 6. Higher rates of P had little influence on yield but significantly increased P removed by the crop.
The clay and clay loam soils of south central and southwestern USA exhibit pH > 7.5 and may contain lime concretions in deeper horizons. Phosphorus activity is lower in soils that have high Ca2+ activity, large amounts of Ca-saturated clay, and highly reactive CaCO3 (Tisdale et al. (1985).) In such soils it is necessary to add larger quantities of P fertilizers to maintain a given level of P activity in the soil solution (Tisdale et al., 1985). However, lacking definitive data for irrigated ryegrass cropping systems, testing laboratories usually recommend no P or relatively small applications of P for these soils. This work was undertaken to provide the basic response curve data for the effects of applied N and P on forage yield and, secondarily, on forage N of irrigated annual ryegrass swards grown on calcareous, high clay soils.
|
MATERIALS AND METHODS
|
|---|
This 3-yr study was conducted on a Knippa clay site (29°18.53' N lat, 99°36.14' W long) that had been used for grazing wheat (Triticum aestivum L.) or ryegrass under center pivot irrigation for >6 yr. Knippa clay typically has a 20-cm dark grayish brown clay Ap horizon containing a few fine, strongly cemented, violently effervescent, CaCO3 concretions <1.0-mm diam. The A horizon at 25 to 45 cm contains brown clay with a few strongly cemented CaCO3 concretions <3 mm in size. The Bw horizon at 45 cm is 25-cm thick with reddish brown clay and similar CaCO3 concretions.
Fertilizer treatments were 0, 135, 269, 404, or 538 kg N ha–1 and 0, 10, 20, or 39 kg P ha–1 applied in a factorial arrangement on plots that were 10 by 4.25 m in a randomized complete block design with four replications. Replications were aligned with the arc of travel of the irrigation boom. A treatment of 20 kg P ha–1 plus sufficient N to maintain tissue N above 32 g kg–1 in leaf tips was applied to an additional plot in each replication (total n = 21). Nitrogen content of leaf tips from forage grown on this treatment was monitored at varying intervals from December through mid-April to determine the need for additional increments (34 kg ha–1) of NA. An examination of results at the end of Year 1 showed no evidence of a plateau in DM yield with increasing P application at the two highest levels of NA. Therefore, plots that were fertilized at the rate of 10 kg P ha–1 in the first year were reassigned to receive 59 kg P ha–1 in the second and third years of the study, and the additional treatment was changed to 538 and 78 kg ha–1 of N and P, respectively, set a priori without regard for potential effects on forage N. Except for the changes noted, treatments were applied repeatedly to the same plots each year.
All of the P and one-third of the N (one-fourth of the N for the 538-kg treatment) were applied before planting and incorporated into the top 10 cm of the soil. For plots receiving 135 to 404 kg N ha–1, remaining increments of N were applied to the surface of each plot in December and February, following the first and second harvests, and, for the 538-kg ha–1 rate of N, a fourth increment was applied in early March. Ammonium nitrate was the source of N, and triple super phosphate was the source of P.
The plot area was cultivated with a disc plow (
14 cm deep) in June and with a sweep cultivator (12 cm deep) in early September each year, lying fallow each summer. Fertilizer was incorporated with a rotary tiller (10 cm deep) in late September. Annual ryegrass (cv TAM-90) was planted about October 1 in 18-cm drill rows at the rate of 34 kg seed ha–1 (Evers et al., 1997). Plots were irrigated (3 cm) within 3 d after planting unless rainfall was adequate for germination and seedling survival. Additional irrigations were applied when dulling of leaf surfaces in late afternoon indicated moisture stress. A gauge to measure precipitation and irrigation amounts was placed in each replication.
Forage was harvested six times each year between mid-December and the end of May. All plots were cut with a self-propelled harvester (Almaco, Nevada, IA) whenever estimated harvestable yield of the most productive plots exceeded 1300 kg ha–1. There was no harvest in January, but there were two harvests each year in April or May. Cutter bar height was set at 8 cm. For yield measurement, a single swath, 1.31 m wide, was harvested from the center of each plot. Length of each harvest swath (8 m) was recorded. Plot yields were weighed using the electronic scale of the harvester's forage collection box. A sample of forage (300 g) from each plot was sealed in a bag, placed on ice immediately after collection, and transported to the laboratory, where it was oven dried to constant weight on an open pan at 105°C (1.5 h) to determine DM content. Dried samples were retained for later processing and analysis. All remaining growth above 8 cm was removed from the plot site immediately after each harvest.
Dried forage samples were ground to pass a 1-mm screen. In each year, numbers 1 to 21 were arbitrarily assigned to the 21 treatment combinations. For each harvest, samples from 14 treatments were composited across replications to reduce the analytical work load, and samples from the remaining seven treatments were tested on an individual plot basis to allow evaluation of variation among plots within treatment. Treatments numbered 1 to 14 were composited for Harvests 1 and 4; those numbered 1 to 7 and 15 to 21 were composited for Harvests 2 and 5; those numbered 8 to 21 were composited for Harvests 3 and 6.
Nitrogen in forage was determined with a micro-Kjeldahl procedure (Tecator System II). For measurement of P concentration, forage samples were digested in nitric acid (Havlin and Soltanpour, 1989) at 105°C. Forage digests were analyzed for P by the vanadomolybdic acid method (Jackson, 1958).
Soil samples were collected from each plot before initial fertilizer application and planting and again within 2 wk after final forage harvest for each of three growth seasons. Two soil cores (4-cm diam.) were taken to a depth of 90 cm in each plot and composited within four soil depths, 0 to 15, 15 to 30, 30 to 60, and 60 to 90 cm. The sample for the surface layer was further composited with soil collected by hand probe from 12 random locations within each plot. All samples were held at 4°C until they were placed in open pans and dried in a forced air oven at 60°C. Soil nitrate N was determined by Cd reduction (Keeney and Nelson, 1982). Plant-available soil P, K, Ca, Mg, Na, and S were determined using Mehlich-3 extraction (Mehlich, 1984), analyzed by inductively coupled plasma–optical emission spectrophotometry. Soil pH was determined in a 1:2 soil/water extract of the soil using deionized water.
Individual plot data describing ryegrass responses to N and P were examined with the general linear model procedure of SAS Institute (1999) and, where appropriate, the more specialized procedure for response surfaces, SAS's RSREG. Sources of variation in the statistical model for DM yield, P uptake, and N uptake were year (Yr), replication, NA (kg ha–1), PA (kg ha–1), initial soil NS (mg kg–1), and initial estimated PS (mg kg–1). The quadratic components, NA x NA, PA x PA, NS x NS, PS x PS, and interactions, NA x PA, NA x PS, NA x NS, PA x PS, PA x NS, NS x PS, Yr x NA, Yr x PA, and Yr x NA x PA were also examined. The variables, year and replication, were treated as discrete while all other variables were regarded as continuous. Relative importance of statistically significant components of the model was judged from the magnitude of the absolute values of t associated with the coefficients estimated for those components.
|
RESULTS AND DISCUSSION
|
|---|
At the initiation of the experiment, mean pH was 8.1 and mean concentration of nitrate N was 6 mg kg–1 in the 15-cm surface layer of soil. Mean plant-available mineral concentrations were P, 25 mg kg–1; K, 664 mg kg–1; Ca, 31 g kg–1; Mg, 364 mg kg–1; S, 615 mg kg–1; and Na, 61 mg kg–1.
Rainfall totals from planting to final harvest were 155, 380, and 278 mm, and total irrigation amounts were 341, 138, and 235 mm for the three seasons, respectively. Rain provided adequate water for almost the entire period of rapid growth during Year 2. No differences (P > 0.8) were noted among rain gauges on different replications.
Dry Matter Yield
For DM yield, the two-way interactions, Yr x NA and Yr x PA, were significant (P < 0.01), indicating the need to examine years independently. Figures 1
, 2
, and 3
show treatment means for DM yield plotted as surface graphs, one for each year of this experiment. For the second and third years, responses to the combination of 538 kg NA ha–1 and 78 kg PA ha–1 are not shown in the graphs, but data for those plots were included in the derivation of the model coefficients in the legends of Fig. 2 and 3. For each of the 3 yr, both the linear and quadratic components of NA and PA, as well as their interaction, were significant factors (P < 0.01) in the equations that modeled DM yields.
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 1. Treatment means for DM yield, Year 1. Model equation: t ha–1 = 2.836 + (NA x 1.142 x 10–2) + (PA x 3.907 x 10–2) + (NA x PA x 1.435 x 10–4) – (NA x NA x 1.585 x 10–5) – (PA x PA x 6.399 x 10–4), (n = 83, R2 = 0.92, s = 0.419).
|
|
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 2. Treatment means for DM yield, Year 2. Model equation: t ha–1 = 1.071 + (NA x 1.959 x 10–2) + (PA x.129) + (NA x PA x 1.746 x 10–4) – (NA x NA x 3.079 x 10–5) – (PA x PA x 1.690 x 10–3), (n = 83, R2 = 0.93, s = 0.764).
|
|
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 3. Treatment means for DM yield, Year 3. Model equation: t ha–1 = 0.964 + (NA x 1.330 x 10–2) + (PA x 0.1448) + (NA x PA x 1.555 x 10–4) – (NA x NA x 2.348 x 10–5) – (PA x PA x 1.874 x 10–3), (n = 83, R2 = 0.89, s = 0.846).
|
|
In Year 1 (Fig. 1), there was no evidence of a plateau in DM yield with increasing P application at the two highest levels of applied N. Analysis of yield data for Year 1 with RSREG (SAS Institute, 1999) projected values of NA and PA associated with maximum yield (11.4 t ha–1) to be 1112 and 144 kg ha–1, respectively, well beyond the experimental range.
Dry matter yield responses to the first increments of NA and PA were much greater in Year 2 than in Year 1, as can be seen from a comparison of Fig. 1 and 2 and from the regression coefficients on NA and PA in the model equations. Model projections for maximum yield were 10.1 and 8.36 t ha–1 in Years 2 and 3, lower than in Year 1 but with associated predictions for required NA and PA within the range of this experiment. Contrary to the inference from the first year's data, levels of PA above 39 kg ha–1 provided no significant (P > 0.05) increase in DM yield at any level of NA in the second or third (Fig. 2 and 3) years.
In Year 3, significant (P < 0.05) DM responses to NA were limited to the first increment except at 59 kg PA ha–1, where there was a further response to the second increment of NA. Growth response to PA extended to the second increment (39 kg ha–1) at all levels of NA except 0 kg NA ha–1. In fitting the full model to DM responses for Year 3, a number of significant (P < 0.05) interactions involving NS and PS were found, although all of them combined accounted for only 2% of the variation in yield, and NS and PS as linear components were themselves not significant (P > 0.05). Furthermore, all significant interactions involving NS and PS had regression coefficients that were opposite in sign to the coefficients of equivalent interactions for NA and PA. Dropping nonsignificant interactions (P > 0.05) from the model produced a stepwise loss and weakening of the remaining interactions, further demonstrating the volatile and spurious nature of NS- and PS-based interactions, which were fed by the significant (P < 0.01) lack of fit (PROC RSREG; SAS Institute, 1999) of the mathematical forms of the model to the shape of the response surface (Fig. 3). Nevertheless, the coefficient of determination was 0.89 for Year 3 compared to 0.92 and 0.93 for Years 1 and 2, respectively.
Concentration of available P in the surface 15 cm of soil (PS = 25 mg kg–1, s = 2.9) at the beginning of the experiment was sufficient to allow a response of almost 2 t of forage ha–1 (P < 0.05) to the highest levels of NA (Fig. 1) in the P control plots. Forage yield response to NA in the plots that received no PA declined in the second season (Fig. 2) and disappeared entirely (P = 0.53) (Fig. 3) in the third season. The diminished response in these plots did not reflect levels of PS, however, which increased and averaged 35 (s = 3.4) and 30 mg kg–1 (s = 4.0) at the beginning of the second and third seasons, respectively, with no differences among plots (P > 0.60) assigned to different levels of NA.
In the N control plots, DM yield did not respond significantly to PA in any year. Mean NS in the surface 15 cm of control plots increased from 3 mg kg–1 before the first season to 7 mg kg–1 at the beginning of the second season and decreased again to 2 mg kg–1 before the third season. The changes in NS coincided with low rainfall during the summer fallow before the second ryegrass season and high rainfall (>15 cm) in June before the third season, which spawned a heavy growth of false ragweed (Parthenium hysterophorus L.).
The curvilinear responses to increasing NA and PA in this experiment, particularly in Years 2 and 3, have been frequently observed (Allen et al., 1974; Dunivan, 1975; Matocha, 1972; Morris et al., 1994; Robinson et al., 1987) and were expected. In the case of P, the quadratic effect was most pronounced in Years 2 and 3, when yield responses to 0 PA at higher levels of NA were notably lower than in Year 1. The cropping history of the plot area and the data from this experiment, when placed in the context of the model used by Jones et al. (1984), suggest that P coming from organic pools was important for 0 PA plots in Year 1. This source was poorly replenished under our harvest regime, and P uptake fell to 5 kg ha–1 for 0 PA plots in Year 3. The commonly observed curvilinear response to P fertilizer may have been further disguised in Year 1 by an increased sequestration of P into active and stable inorganic pools. No P had been applied to the site for at least 6 yr before initiation of this experiment. For 20 and 39 PA plots that also received >0 NA, average net efficiency of P uptake increased from 23% in Year 1 to more than 53% in Years 2 and 3.
Curvilinear growth responses to increasing levels of soil nutrients are inherently better modeled by the exponential form, Y = a x (1 – e–kX) (where Y is DM yield, a and k are constants, and X is NA or PA), than by the multivariate linear model that was used for the model equations, but its advantage is restricted to two-dimensional applications. The exponential model demonstrated exceptionally good fit to two-dimensional curvilinear data. However, in a forced three-dimensional application [i.e., Z = a x (1 – e–kXY) + C] (Z is DM yield, a and k are constants, X and Y are NA and PA, and C is an additive constant), it provided no better fit overall than the multivariate model. Furthermore, it does not readily allow statistical evaluation of the importance of potential independent variables in growth response.
The change in PA treatment levels between Years 1 and 2 allowed us to confirm that 40 kg PA ha–1 approximated the economic optimum level for production of irrigated ryegrass on calcareous clay soils in this region. Projected levels of PA for maximum production in Years 2 and 3 were 64 and 58 kg ha–1, respectively. Economically optimal production frequently occurs in the range of 80 to 90% of maximum production (J.G. Pena, personal communication, 2005). Model predictions for PA required for this range of production are 31 to 41 kg ha–1. The adequacy of this level of PA agrees reasonably well with data from Robinson and Eilers (1996), who stated that 45 to 50 kg PA ha–1 produced 90% of maximum ryegrass DM yield on a Tangi silt loam at pH 6. Hillard (1989) reported that ryegrass yielded 98% of maximum at 30 kg PA ha–1 on well-limed, Lilbert loamy fine sand with essentially no extractable Al. Application of 39 kg P ha–1 in this experiment resulted in PS means of 40 and 32 mg kg–1 at the beginning of Years 2 and 3, respectively, somewhat higher than PS for plots with 0 PA (35 and 30 mg kg–1), with no evidence of increase with time.
Levels of NA predicted from model projections for maximum DM yield, were 499 and 477 kg ha–1 for Years 2 and 3, respectively. Levels of NA for 80 to 90% of maximum production were much lower, ranging from 250 to 315 kg ha–1 for Years 2 and 3. Furthermore, amounts of NA above 269 kg ha–1 provided no significant (P > 0.05) growth response at any level of PA in any year of this experiment. This NA rate is considerably lower than N rates of 450 kg ha–1 reported to produce highest statistically significant yields of annual ryegrass in the South and southeastern states (Dunivan, 1975; Robinson et al., 1987).
Phosphorus Uptake
Phosphorus uptake was significantly (P < 0.05) related to Yr and to Yr x NA x PA. Therefore, treatment means for P uptake were presented within year (Fig. 4
, 5
, 6
). The interaction of years with NA and PA can be seen in the differing shapes of the surface graphs and the magnitude of the regression coefficients that model them. The regression coefficients on NA and PA were much larger for both linear and quadratic aspects, in absolute terms, in the second year. For NA the coefficients declined again in Year 3, but remained elevated, compared to the first year, for PA.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 4. Treatment means for P uptake, Year 1. Model equation: kg ha–1 = 7.034 + (NA x 0.02918) + (PA x 0.1529) + (NA x PA x 3.671 x 10–4) – (NA x NA x 4.18 x 10–5) – (PA x PA x 2.097 x 10–3), (n = 20, R2 = 0.98, s = 0.69).
|
|
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 5. Treatment means for P uptake, Year 2. Model equation: kg ha–1 = 2.809 + (NA x 0.05415) + (PA x 0.4300) + (NA x PA x 6.046 x 10–4) – (NA x NA x 8.661 x 10–5) – (PA x PA x 5.315 x 10–3), (n = 21, R2 = 0.97, s = 2.05).
|
|
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 6. Treatment means for P uptake, Year 3. Model equation: kg ha–1 = 2.998 + (NA x 0.02890) + (PA x 0.3777) + (NA x PA x 3.863 x 10–4) – (NA x NA x 5.246 x 10–5) – (PA x PA x 4.280 x 10–3), (n = 21, R2 = 0.93, s = 2.10).
|
|
In each year, P uptake was significantly related (P < 0.01) to NA and PA (both linear and quadratic components) and to NA x PA, but not to NS, PS, or any of their interactions. As expected, the surface graph for P uptake in each year has a shape similar to the respective DM yield response. As with DM yield (Fig. 1), Fig. 4 shows little evidence of a plateau in P uptake as PA increases at the higher rates of NA during the first year; PA associated with the projected maximum P uptake was 109 kg ha–1, far outside the range of treatments in Year 1. Phosphorus concentration in ryegrass forage increased (P < 0.01) with increasing levels of PA, but it was not influenced by NA.
For Harvests 3 through 6 in Year 2, the concentration of P in forage was significantly greater (P < 0.01) than in either of the other years (Fig. 7
). As indicated by the magnitude of t values, increased concentration of P in forage was the major factor contributing to increased P uptake in Year 2. The pattern of P concentration in forage was similar in each year (Fig. 7) and appeared to follow the patterns of ambient temperature or daylength, although physiological maturity of the plants may also have had an influence.
Nitrogen Uptake
In Year 1, 269 kg N ha–1 were required for the extra plots to meet the criterion of maintaining N in leaf tips >32 g kg–1; their mean DM yield (6130 kg ha–1) was similar to the yield from those plots designated a priori to receive 269 kg N and 20 kg P ha–1 (5980 kg ha–1). The data in Table 1 show that 269 kg NA ha–1 were generally associated with sufficient forage N to meet the 20% CP quality criterion. Nitrogen uptake was significantly related (P < 0.01) to Yr and to Yr x NA x PA. Means for N uptake in individual years are plotted in Fig. 8
, 9
, and 10
. The linear and quadratic components of NA as well as the interaction of NA and PA, strongly influenced (P < 0.01) N uptake in all 3 yr. Linear and quadratic components of PA were also strong factors associated with N uptake in Years 2 and 3. However, because the influence of PA on N uptake was effective only through forage yield, PA was statistically weak (P < 0.10) as a factor in N uptake in the first year. No effect of PA on forage N content was observed (Table 1). On the other hand, NA had a very strong effect on forage N (P < 0.01) (Table 1) and, therefore, had a positive influence on both components of the arithmetic product, N uptake. Note, also, that in the model equations for the means plotted in Fig. 8
to 10, the regression coefficient for NA is greater (P < 0.01) in the second year than in either of the other years.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 8. Treatment means for N uptake, Year 1. Model equation: kg ha–1 = 40.16 + (NA x 0.5673) + (PA x 1.314) + (NA x PA x 4.332 x 10–3) – (NA x NA x 6.625 x 10–4) – (PA x PA x 2.236 x 10–2), (n = 20, R2 = 0.99, s = 7.70).
|
|
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 9. Treatment means for N uptake, Year 2. Model equation: kg ha–1 = 5.62 + (NA x 0.7608) + (PA x 4.234) + (NA x PA x 7.386 x 10–3) – (NA x NA x 10.86 x 10–4) – (PA x PA x 6.043 x 10–2), (n = 21, R2 = 0.97, s = 21.37).
|
|
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 10. Treatment means for N uptake, Year 3. Model equation: kg ha–1 = –4.78 + (NA x 0.5999) + (PA x 4.067) + (NA x PA x 6.418 x 10–3) – (NA x NA x 9.23 x 10–4) – (PA x PA x 5.812 x 10–2), (n = 21, R2 = 0.96, s = 20.62).
|
|
Data from this study suggest that incrementally applying N fertilizer to ryegrass in its vegetative stages of growth to maintain N in leaf tissue 32 g kg–1 may provide economically optimal growth for both the crop and the young cattle grazing it. Such a guide for NA would automatically take into account within-season N recycling in a grazing system.
General Comments
The finding that NS and PS never rose to even minor statistical importance for plant responses was not surprising. In Year 1, the values for NS and PS were essentially the same in all plots and could not correlate with differences in yield. In the second and third years, any change in NS and PS would have been a consequence of application rates. Such changes would have been correlated to application rates and would, therefore, have been statistically shadowed by the continuation of those same application rates.
The basis for substantially greater DM yields and higher uptakes of N and P in Year 2 is not immediately apparent from examination of amounts of rainfall plus irrigation (496, 518, and 513 mm in Years 1, 2, and 3, respectively). However, abnormally high rainfall in late winter and early spring of Year 2 dampened the soil profile at least twice as deep as a nominal irrigation (20 cm). Furthermore, supplemental water was generally not applied until the crop evidenced a slight wilting in late afternoon. Alleviation of short periods of moisture stress and enabling roots to gain nutrients from a larger volume of soil during the period of most rapid plant growth were undoubtedly factors in the higher DM yields in Year 2. More liberal irrigation criteria (e.g., 75% of potential evapotranspiration) with application rates that will dampen most of the nominal root zone (perhaps in rapid succession split applications) may be in order.
|
CONCLUSIONS
|
|---|
Maximum annual ryegrass yield under the conditions of this experiment was 10 t ha–1. Assuming that economically optimal yield occurs in the range of 80 to 90% of maximum yield, predicted NA requirement for economically optimal yield ranges from 250 to 315 kg ha–1. Predicted PA requirement for economically optimal yield ranges from 31 to 41 kg ha–1. Recommended amounts are 270 kg NA ha–1 and 36 kg PA ha–1 under the conditions of this experiment and a 10% increase under less limiting irrigation criteria. Applying increments of N to maintain forage N in leaf tips 32 g kg–1 in vegetative ryegrass will approximate the recommended level of N fertilization while accounting for within-season N recycling under grazing conditions.
|
NOTES
|
|---|
Contribution of the Texas Agric Exp. Stn. and partially supported by a grant from the Potash & Phosphorus Institute.
|
REFERENCES
|
|---|
- Allen, M., H.D. Ellzey, B.D. Nelson, C.R. Montgomery, and P.E. Schilling. 1974. The response of an oat–ryegrass mixture to applied nitrogen, phosphorus, and potassium. Louisiana Agric. Exp. Stn. Bull. 680. Louisiana Agric. Exp. Stn., Baton Rouge.
- Beever, D.E., S.B. Cammell, and A. Wallace. 1974. The digestion of fresh, frozen and dried perennial ryegrass. Proc. Nutr. Soc. 33:73A–74A.[Medline]
- Dunivan, L.S. 1975. Production of rye and ryegrass forage with sulfur-coated urea and ammonium nitrate. Agron. J. 67:415–417.[Abstract/Free Full Text]
- Evers, G.W., G.R. Smith, and C.S. Hoveland. 1997. Ecology and production of annual ryegrass. p. 29–43. In F.M. Rouquette, Jr. and L.R. Nelson (ed.) Ecology, production, and management of Lolium for forage in the USA. CSSA Spec. Publ. 24. CSSA, Madison, WI.
- Haby, V.A., and D.L. Robinson. 1997. Soil fertility and liming practices for production of annual ryegrass. p. 45–69. In F.M. Rouquette, Jr., and L.R. Nelson (ed.) Ecology, production, and management of Lolium for forage in the USA. CSSA Spec. Publ. 24. CSSA, Madison, WI.
- Havlin, J.L., and P.N. Soltanpour. 1989. A nitric acid plant digest method for use with inductively coupled spectrometry. Commun. Soil Sci. Plant Anal. 14:969–980.
- Hillard, J.B. 1989. Coastal bermudagrass and Marshall ryegrass response to surface-applied limestone and phosphorus on an acid, sandy East Texas soil. Ph.D. diss. Texas A&M Univ., College Station.
- Hillard, J.B., V.A. Haby, and F.M. Hons. 1992. Annual ryegrass response to limestone and phosphorus on an Ultisol. J. Plant Nutr. 15:1253–1268.
- Hons, F.M., L.A. Larson-Vollmer, and M.A. Locke. 1990. NH4OAc–EDTA extractable phosphorus as a soil test procedure. Soil Sci. 149:249–256.
- Hovermale, C.H. 1993. Effect of rate and date of ammonium nitrate on yield of ryegrass. Dep. Info. Serv., Div. Agric., For., and Vet. Med., Mississippi State Univ., Mississippi State.
- Jackson, M.L. 1958. Soil chemical analysis. Prentice Hall, New York.
- Jones, C.A., C.V. Cole, A.N. Sharpley, and J.R. Williams. 1984. A simplified soil and plant phosphorus model: I. Documentation. Soil Sci. Soc. Am. J. 48:800–805.[Abstract/Free Full Text]
- Keeney, D.R., and D.W. Nelson. 1982. Nitrogen—inorganic forms. p. 643–687. In A.L. Page et al. (ed.) Methods of soil analysis. Part 2. 2nd ed. Agron. Monogr. 9. ASA and SSSA, Madison, WI.
- Lippke, H., T.D.A. Forbes, and W.C. Ellis. 2000. Effect of supplements on growth and forage intake by stocker steers grazing wheat pasture. J. Anim. Sci. 78:1625–1635.[Abstract/Free Full Text]
- Lippke, H., and D.B. Herd. 1990. FORAGVAL: Intake and gain by cattle estimated from acid detergent fiber and crude protein of the forage diet. Publ. MP-1708. Texas Agric. Exp. Stn., College Station, TX.
- Matocha, J.E. 1972. Production of wheat–ryegrass and Coastal bermudagrass grown in association as affected by rates and sources of nitrogen. Texas Agric. Exp. Stn. Progr. Rep. 3015. Texas Agric. Exp. Stn., College Station.
- Mehlich, A. 1984. Mehlich-3 soil test extractant: A modification of Mehlich-2 extractant. Commun. Soil Sci. Plant Anal. 15:1409–1416.
- Morris, D.R., T.F. Brown, V.C. Baligar, D.L. Corkern, L.K. Zeringue, and L.F. Mason. 1994. Ryegrass forage yield and quality response to sulfur and nitrogen fertilizer on Coastal Plain soil. Commun. Soil Sci. Plant Anal. 25:3035–3046.
- Robinson, D.L., K.G. Wheat, and N.L. Hubbert. 1987. Nitrogen fertilization influences on Gulf ryegrass yields, quality, and nitrogen recovery from Olivier silt loam soil. Louisiana Agric. Exp. Stn. Bull. 784.
- Robinson, D.L., and T.L. Eilers. 1996. Phosphorus and potassium influences on annual ryegrass production. Louisiana Agric. 39(2):10–11.
- SAS Institute. 1999. Release 8.01. SAS Inst., Cary, NC.
- Tisdale, S.L., W.L. Nelson, and J.D. Beaton (ed.). 1985. Soil and fertilizer phosphorus. Soil fertility and fertilizers. 4th ed. Macmillan Publ. Co., New York.
- Westfall, D.G., A.W. Smith, and G.W. Evers. 1971. Effect of urea and sulfur coated urea on the dry matter and protein production of gulf ryegrass. Publ. PR-2906. Texas Agric. Exp. Stn., College Station.
- Worrell, M.A., D.J. Undersander, C.E. Thompson, and W.C. Bridges, Jr. 1990. Effects of time of season and cottonseed meal and lasalocid supplementation on steers grazing rye pastures. J. Anim. Sci. 68:1151–1157.
TOP
ABSTRACT
INTRODUCTION
