Agronomy Journal 93:891-895 (2001)
© 2001 American Society of Agronomy
FORAGES
Residue Removal and Nitrogen Fertilization Affects Tiller Development and Flowering in Meadow Bromegrass
Heather A. Loeppky* and
Bruce E. Coulman
Agriculture and Agri-Food Canada (AAFC), 107 Science Place, Saskatoon, SK, S7N 0X2 Canada
* Corresponding author (heather.loeppky{at}gov.ab.ca)
Received for publication April 10, 2000.
 |
ABSTRACT
|
|---|
Flowering and seed yield in many temperate grasses are dependent on floral induction in the previous fall. Field experiments were conducted in Saskatchewan to determine the effect of crop residue removal and N fertilization on tiller and panicle development in meadow bromegrass (Bromus riparius Rehm.). Four residue removal treatments (none, after harvest, October, and after harvest + October) and three N treatments (0, 50, and 100 kg N ha-1) were applied in each of 2 yr. Tiller density and leaf stage were determined in fall and spring; panicle density was determined just before seed harvest each year. Removing crop residue generally increased tiller and panicle density. However, fall tiller density decreased at Saskatoon in 1996 due to dry conditions, regardless of residue removal. Because fewer tillers were present in the fall when conditions that promote flowering prevailed, panicle density was reduced by 62% compared with the previous year. Nitrogen generally did not affect tiller density or development. However, in the spring of 1996, 100 kg N ha-1 with a single residue removal increased leaf stage from 2.4 to 2.6 leaves tiller-1. This rate of N with double residue removal reduced leaf stage to 2.2 leaves tiller-1 due to winter injury. Fall tiller density and panicle production were similarly affected. As a result of winter injury and drought, fall tiller density and development were not highly or frequently correlated with panicle or seed production. Hence, fall tiller density and development in the prior year cannot be used as a tool to predict seed yield.
|
INTRODUCTION
|
|---|
SEED PRODUCTION in many temperate perennial grasses is closely related to fall tiller development (Schoberlein, 1987; Boelt, 1999). This is particularly important in grasses with a dual floral induction system, including smooth bromegrass (B. inermis Leyss) (Heide, 1994). In such species, tillers must not only be present under conditions favorable to primary induction, but they must also be inducible. Whether the tillers are inducible or not often depends on whether they have gone through the transition from juvenile, nonresponsive growth to mature, responsive growth. The physiological basis for maturity is not well understood, but one of the most accurate morphologic indicators of maturity is the number of leaves on a tiller. A significant correlation has been reported between the number of leaves tiller-1 in the fall and the percentage of panicles in the following year for orchardgrass (Dactylis glomerata L.), meadow fescue (Festuca pratensis Huds.), timothy (Phleum pratensis L.), Kentucky bluegrass (Poa pratensis L.), red fescue (F. rubra L.) and, to a lesser extent, perennial ryegrass (Lolium perenne L.) (Nordestgaard, 1980; Schoberlein, 1987; Boelt, 1999). Canode and Law (1978) found that large spring tillers were more likely to flower than small spring tillers of smooth bromegrass and crested wheatgrass [Agropyron cristatum (L.) Gaertn.]. Recently, Chastain and Young (1998) found that fall tiller size, based on tiller basal diameter, did not account for the variability in seed yield related to various residue management techniques in Kentucky bluegrass. This effect was particularly noticeable as stands aged. Tiller height was more closely negatively correlated with flowering and seed production than with leaf stage, basal diameter, or biomass production. This may be related to the importance of residue removal in Kentucky bluegrass rather than stage of maturity.
Removing crop residue often increases panicle density in cool-season perennial grasses (Fulkerson, 1980; Thompson and Clarke, 1989; Thompson and Clarke, 1993; Chastain et al., 1997). Residue removal reduced tiller height and increased the percentage of large-sized tillers in Kentucky bluegrass (Chastain et al., 1997). Clipping height did not affect either tiller basal diameter or the number of panicles in Kentucky bluegrass (Thompson and Clarke, 1993).
Nitrogen fertilization and crop residue removal are important for tiller and panicle development as well as for seed production (Canode and Law, 1978; Nordestgaard, 1980; Thompson and Clarke, 1989; Thompson and Clarke, 1993). Nitrogen is required at two critical plant developmental stages: (i) preinduction, to ensure that an adequate number of well-developed tillers are present for induction and (ii) inflorescence elongation, to ensure that adequate N is available to support elongation and floret filling. Nitrogen fertilizer increased tiller density, tiller basal diameter, panicle production, and seed yield in Kentucky bluegrass (Thompson and Clarke, 1989; Thompson and Clarke, 1993). In Holland, delaying spring N application from the end of February (before growth is initiated) until the end of March or April resulted in fewer panicles in red fescue, orchardgrass, meadow fescue, and perennial ryegrass (Meijer and Vreeke, 1988a). However, in all but orchardgrass, seed number tiller-1 and seed mass compensated so that seed yield was not affected by the delay in N application. Floral initiation in orchardgrass occurs very early (January or February); late application of N stimulated vegetative growth at the expense of reproductive growth. Schoberlein et al. (1995) used labelled N to demonstrate that N applied in late summer, autumn, or spring was used mainly for inflorescence emergence in ryegrass, timothy, orchardgrass, Kentucky bluegrass, and meadow fescue.
Positive seed yield response has been observed in a number of grasses, including smooth bromegrass (Meijer and Vreeke, 1988a; Loeppky et al., 1999). Seed yield of smooth bromegrass generally increases more in response to fall-applied N than spring-applied N (Knowles and Cooke, 1952; Buller et al., 1955). However, in certain cases (Loeppky et al., 1999), spring N resulted in higher smooth bromegrass seed yield than fall N application. Higher seed yield of smooth bromegrass following mid-May N application compared with that of mid-April was related to rainfall (Harrison and Crawford, 1941). No information is available on the influence of crop residue removal and N fertilization on tiller growth and development in meadow bromegrass or how fall tiller development is related to flowering in the following year.
The objectives of this study were to determine whether (i) crop residue removal and N fertilization influenced meadow bromegrass tiller growth, development, and fertility and (ii) flowering in meadow bromegrass was related to fall tiller density and development.
|
MATERIALS AND METHODS
|
|---|
Field experiments were conducted under rainfed conditions at Saskatoon, SK, Canada (52°07'N, 106°38'W) and under irrigation at Outlook, SK, Canada (51°30'N, 107°03'W). Trials at both sites were seeded on Typic Haploboroll soils. At Saskatoon, growing season precipitation (April–October) was 307, 282, 366, and 247 mm in 1994, 1995, 1996, and 1997, respectively. At Outlook, growing season precipitation was 216 mm plus an additional 150 mm of irrigation applied in 1994, 272 mm of rainfall plus 225 mm of irrigation in 1995, 298 mm of rainfall plus 106 mm of irrigation in 1996, and 197 mm of rainfall plus 319 mm of irrigation in 1997. Tensiometers were used to determine soil moisture, and irrigation was applied to maintain a consistent soil moisture level of 50% of the available water-holding capacity (Sonmor, 1963).
‘Paddock’ meadow bromegrass was seeded on 10 June 1994 at 12 kg ha-1 in rows spaced 30 cm apart. Fertilizer (11–51–0), to supply 50 kg P ha-1 and 11 kg N ha-1, was incorporated during seedbed preparation. Each fall, an additional 50 kg P ha-1 and 11 kg N ha-1 (11–51–0) was broadcast on all plots. In the fall of 1995, before fertilization of the plot areas, soil cores were taken from each plot at sampling depths of 0 to 15, 15 to 30, and 30 to 60 cm. These samples were air-dried and analyzed for NO3–N and sodium bicarbonate (NaHCO3) soluble P by methods of Gentry and Willis (1988) and Hamm et al. (1970), respectively. Available N levels at 0 to 15, 15 to 30, and 30 to 60 cm were 1.8, 0.9, and 1.8 kg ha-1, respectively, at Outlook and 3.6, 0.9, and 6.2 kg ha-1, respectively, at Saskatoon.
Treatments consisted of 0, 50, and 100 kg N ha-1 and four residue removal treatments: none, after harvest, October, or after harvest + October. These treatments were applied to the same plots each year. Seed was harvested in the last week of July, and appropriate residue treatments were applied within a week of harvest. The October time of removal represents the end of the growing season on the Canadian Prairies. Treatments were replicated four times. The design was a split plot with N rate as the main plot and residue removal as the subplot. Subplots were 1.5 by 6.0 m. Ammonium nitrate (NH4NO3) was broadcast at the prescribed N rates in early to mid-September each year starting in 1994. Residue treatments consisted of removal with a forage harvester and then scalping with a flail mower to a height of 2.5 cm, which removed 90 to 95% of the residue each year starting in 1995. Two 0.25-m2 quadrats were marked in each plot. Starting in the fall of 1995, tiller density and development were determined in early October and mid-May each year. These dates coincide with the beginning and end of the growing season in the area. Methods described in Haun (1973) were modified to stage grass tiller development. Panicles were counted in the same quadrats after anthesis each year.
Analyses of variance were conducted using the general linear model (GLM) procedure of SAS with all effects, except replications, considered fixed effects (SAS Inst., 1985). Data from each year were analyzed separately because a combined analysis showed significant interactions between years and treatments. The data for the two sites were analyzed separately when a significant interaction occurred between site and one or both of the other variables. When interactions were not significant, main effects were presented. The effects of N rates were partitioned into linear and quadratic components, and the means of residue removal treatments were evaluated through a series of a priori orthogonal contrasts. A priori contrasts were also used to determine the nature of interactions between N rate and residue removal treatments. Differences were considered statistically significant at P 
0.05. Standard errors were calculated for all means. Pearson correlation coefficients were calculated on treatment means to determine the relationship between tiller density and development and panicle density and between panicle density, tiller density, or tiller development and seed yield.
|
RESULTS AND DISCUSSION
|
|---|
Tiller density was lower when residue was removed late (October) compared with early (after harvest) at both sites in the fall of 1995 (Table 1). In the spring of 1996, tiller density decreased due to winter injury in treatments where the residue was removed in October. In a previous study, reduced tillering was also attributed to late residue removal (burning) in timothy (Entz et al., 1994). Kentucky bluegrass and creeping red fescue also require a regrowth period after residue removal (Meijer and Vreeke, 1988b). However, in the present study, late removal reduced tiller density in fall and spring of only 1 of the 2 yr.
View this table:
[in this window]
[in a new window]
|
Table 1. The effect of residue removal on tiller density in Paddock meadow bromegrass at Saskatoon and Outlook, SK, Canada 1995–1997.
|
|
By the fall of 1996, tiller densities had increased to more than 3000 m-2 at Outlook in treatments where residue was removed immediately after harvest (Table 1). In Saskatoon, tiller density was much lower than at Outlook and residue removal had no effect. Higher moisture levels at Outlook due to the combination of precipitation and irrigation, combined with mowing, stimulated tillering, whereas the dry fall conditions and lack of rainfall after N fertilization at Saskatoon suppressed tiller development in the fall. In the spring of 1997, tiller numbers were much higher in treatments where residue was removed than where it was left intact, regardless of timing or frequency.
When residue was removed, the developmental stage of tillers was equal to or more advanced than when residue was not removed (Table 2). In the fall of 1996, residue removal increased leaf number tiller-1 at both sites (Table 2). Residue removal also had a positive effect on tiller development at Saskatoon but no effect at Outlook in the spring of 1997. Residue removal increased light penetration through the canopy and changed light quality in other cool-season grasses (Silvertown, 1980; Meijer and Vreeke, 1988b); however, the effect on tiller density and size was variable. Tiller number and size increased when residue was removed in some cases (Thompson and Clarke, 1993), but removal did not consistently advance tiller development in others (Chastain et al., 1997).
View this table:
[in this window]
[in a new window]
|
Table 2. The effect of residue removal on tiller development in Paddock meadow bromegrass at Saskatoon and Outlook, SK, Canada 1995–1997.
|
|
At Outlook in 1996, a significant interaction occurred between N and single vs. double residue removal. Leaf number tiller-1 increased with increasing N rate in the single residue removal treatments but decreased with increasing N rate when the residue was removed twice (Table 3). There was also a significant interaction between N (quadratic) and single vs. double residue removal for panicle density and percentage of tillers that produced panicles (Table 4). Panicle production increased linearly with N rate when residue was removed after harvest, whereas when residue was removed twice, panicle production increased when 50 kg ha-1 was applied but declined when 100 kg ha-1 was applied. Both clipping and N application decrease cold hardiness of cool-season grasses by reducing total soluble carbohydrates and increasing tissue hydration (Huokona, 1974; Smith, 1987). In this case, double residue removal alone did not reduce tiller development or panicle production; however, the combination of high N rate and double residue removal negatively affected both tiller and panicle development at Outlook. Under drier conditions at Saskatoon, winter injury did not occur even with double residue removal and high N rates.
View this table:
[in this window]
[in a new window]
|
Table 3. The effect of residue removal and N fertilization on tiller development in Paddock meadow bromegrass at Saskatoon and Outlook, SK, Canada, May 1996.
|
|
View this table:
[in this window]
[in a new window]
|
Table 4. The effect of residue removal and N fertilization on panicle density and percentage (based on previous fall tiller density) in Paddock meadow bromegrass for combined sites in 1996.
|
|
In 1997, panicle density increased significantly with increasing N rate (256, 363, and 375 panicles m-2 with 0, 50, and 100 kg N ha-1, respectively) and residue removal. More panicles were produced when residue was removed early than late (350 vs. 76 m-2, respectively). A second removal increased panicle density compared with the mean of the two early removal treatments (322 vs. 229 m-2, respectively).
The effect of residue removal on the percentage of fall tillers that produced panicles varied between sites in both years (Table 5). At Saskatoon in 1996, a significant increase in tiller fertility occurred when residue was removed compared with the no-removal treatment. Within the removal treatments, double removal resulted in a greater proportion of panicles than single removal. At Outlook in 1996, a higher percentage of panicle-producing tillers occurred in early removal compared with late-removal treatments. This was attributed to winter injury, which resulted in a decrease in tiller density from the fall of 1995 to the spring of 1996 in plots where residue was removed in October (Table 1). In 1997, the percentage of panicles had decreased, but it was higher when residue was removed than when it was not removed at both sites. At Saskatoon, double removal resulted in a higher percentage of panicles than single removal, and early residue removal was higher than late residue removal. Residue removal has previously been associated with increased panicle production in Kentucky bluegrass (Thompson and Clarke, 1989; Chastain et al., 1997) and creeping red fescue (Meijer and Vreeke, 1988b).
View this table:
[in this window]
[in a new window]
|
Table 5. The effect of residue removal on the percentage of tillers with panicles (based on fall tiller density) in Paddock meadow bromegrass at Saskatoon and Outlook, SK, Canada 1996 and 1997.
|
|
Fall tiller growth and development are of particular interest to those involved in seed production of temperate perennial grasses because flowering and seed yield are correlated with fall tiller growth in many species (McDonald and Copeland, 1996; Schoberlein, 1987; Boelt, 1999) although this is not always the case (Meijer and Vreeke, 1988b). In our study, correlations with panicle density were high and significant in 1997 for fall tiller density at Outlook and for spring density at both sites but were otherwise nonsignificant (Table 6). Correlations between fall or spring tiller density and seed yield were lower and often nonsignificant. Spring tiller development was more closely correlated to panicle density and seed production in the following season than fall or spring tiller density. Fall tiller development correlations with panicle density and seed yield were small and nonsignificant.
View this table:
[in this window]
[in a new window]
|
Table 6. Pearson correlation coefficients (r) between panicle density or seed yield and tiller density and stage of tiller development (leaves tiller-1) in Paddock meadow bromegrass at Saskatoon and Outlook, SK, Canada 1996 and 1997 (n = 12).
|
|
Although Schoberlein (1987) found a significant correlation between leaf number tiller-1 in fall and flowering tillers in the following year for orchardgrass, timothy, and meadow fescue, he also found that flowering occurred once a certain leaf number had been reached. Schoberlein's findings support the theory that a juvenile period exists in many perennial plants during which flowering will not occur even if the plant is exposed to conditions that stimulate flowering at a later growth stage (Heide, 1994). However, Havstad (1996) has presented evidence that tillers, which emerge during induction conditions, are capable of becoming reproductive in Scandinavian cultivars of smooth bromegrass, Kentucky bluegrass, meadow fescue, orchardgrass and ryegrass. Growth chamber studies indicate that this is also the case with meadow bromegrass (Loeppky, 1999). In the present study, panicle density was the variable most closely correlated with seed yield at both sites in both years (Table 6). This is also the case with Kentucky bluegrass and creeping red fescue (Meijer and Vreeke, 1988b). Meijer and Vreeke (1988a) speculate that competition for light and nutrients during elongation, rather than the number of inducible tillers, limits heading. In this study, weather conditions between fall and flowering in the following year (drought and winter injury), in combination with certain treatments, had a major effect on tiller growth and development.
|
CONCLUSIONS
|
|---|
Residue removal immediately after harvest generally increased tiller density and development as well as panicle density and the percentage of panicle-producing tillers in meadow bromegrass. However, good moisture conditions combined with heavy clipping and high N rates resulted in winter injury during a severe winter. This reduced tiller density, stage of development, panicle density, and the percentage of panicles. Fall drought resulted in low fall tiller density and reduced the percentage of panicles in the following season at one location. This indicates that different management practices are required under different moisture conditions. However, removing residue immediately after harvest consistently increased panicle production. The correlation between fall tiller density and development and panicle or seed production, while statistically significant, was low and, as such, cannot be used as a tool to predict whether or not seed production will be economically viable in the following year.
|
ACKNOWLEDGMENTS
|
|---|
The authors would like to thank Bruce Hesselink, Don Penner, and Cheryl Duncan for their technical assistance on this study.
|
NOTES
|
|---|
H.A. Loeppky, current address: Alberta Agriculture, Food, and Rural Development, 6000 C&E Trail Lacombe, AB, T4L 1V5 Canada.
|
REFERENCES
|
|---|
- Boelt, B. 1999. The effect of tiller size in autumn on the percentage of reproductive tillers in amenity types of Poa pratensis L., Festuca rubra L., and Lolium perenne L. p. 53–57. In Proc. Int. Herbage Seed Conf., 4th, Perugia, Italy. 23–27 May 1999. University of Perugia, Italy.
- Buller, R.E., J.S. Bubar, H.R. Fortmann, and H.L. Carnahan. 1955. Effects of nitrogen fertilization and rate and method of seeding on grass yields in Pennsylvania. Agron. J. 47:559–563.[Free Full Text]
- Canode, C.L., and A.G. Law. 1978. Influence of fertilizer and residue management on grass seed production. Agron. J. 70:543–546.[Abstract/Free Full Text]
- Chastain, T.G., G.L. Kiemnec, G.H. Cook, C.J. Garbacik, B.M. Quebbeman, and F.J. Crowe. 1997. Residue management strategies for Kentucky bluegrass seed production. Crop Sci. 37:1836–1840.[Abstract/Free Full Text]
- Chastain, T.G., and W.C. Young III. 1998. Vegetative plant development and seed production in cool-season perennial grasses. Seed Sci. Res. 8:295–301.
- Entz, M.H., S.R.J. Smith, D.J. Catani, and A.K. Storgaard. 1994. Influence of post-harvest residue management and cultivar on tiller dynamics and seed yield in timothy. Can. J. Plant Sci. 74:507–513.
- Fulkerson, R.S. 1980. Seed yield response of three grasses to post-harvest stubble removal. Can. J. Plant Sci. 60:841–846.
- Gentry, C.E., and R.B. Willis. 1988. Improved method for automated determination of ammonium in soil extracts. Commun. Soil Sci. Plant Anal. 19:721–737.
- Hamm, J.W., F.G. Radford, and E.H. Halstead. 1970. The simultaneous determination of nitrogen, phosphorus, and potassium in sodium bicarbonate extracts of soils. p. 65–79. In Advances in automated analysis. Vol. 2. Industrial analyses. Proc Technicon Int. Congr.
- Harrison, C.M., and W.N. Crawford. 1941. Seed production of smooth bromegrass as influenced by application of nitrogen. J. Am. Soc. Agron. 33:643–651.
- Haun, J.R. 1973. Visual quantification of wheat development. Agron. J. 65:116–119.[Abstract/Free Full Text]
- Havstad, L.T. 1996. Juvenility and flowering in Festuca pratensis Huds: 1. Effects of plant age, cultivar, and duration of primary induction treatments. Norw. J. Agric. Sci. 10:159–178.
- Heide, O.M. 1994. Control of flowering and reproduction in temperate grasses. New Phytol. 128:347–362.
- Huokona, E. 1974. Wintering of heavily fertilized grasslands. Proc. Int. Grassl. Congr. 2:213–218.
- Knowles, R.P., and D.A. Cooke. 1952. Response of bromegrass to nitrogen fertilizers. Sci. Agric. (Ottawa) 32:548–554.
- Loeppky, H.A. 1999. Flowering and seed production in meadow bromegrass. Ph.D. diss. Univ. of Saskatchewan, Saskatoon, SK, Canada (Diss. Abstr. AATNQ43515).
- Loeppky, H.A., P.R. Horton, S. Bittman, L. Townley-Smith, T. Wright, and W.F. Nuttall. 1999. Forage seed yield response to N and P fertilizers and soil nutrients in northeastern Saskatchewan. Can. J. Soil Sci. 79:265–271.
- McDonald, M.B., and L.O. Copeland. 1996. Seed production principles and practices. Chapman and Hall, New York.
- Meijer, W.J.M., and S. Vreeke. 1988a. Nitrogen fertilization of grass seed crops as related to soil mineral nitrogen. Neth. J. Agric. Sci. 36:375–385.
- Meijer, W.J.M., and S. Vreeke. 1988b. The influence of autumn cutting treatments on canopy structure and seed production of first-year crops of Poa pratensis L. and Festuca rubra L. Neth. J. Agric. Sci. 36:315–325.
- Nordestgaard, A. 1980. The effects of quantity of nitrogen, date of application, and the influence of autumn treatment on the seed yield of grasses. p. 105–119. In P.D. Hebblethwaite (ed.) Seed production. Butterworth, London, England.
- SAS Institute. 1985. SAS user's guide: Statistics. Version 5. SAS Inst., Cary, NC.
- Schoberlein, W. 1987. Correlations between phase of development of some perennial grass species in autumn and seed yield characteristics in the following year. J. Appl. Seed Prod. 5:60 (abstr.
- Schoberlein, W., M. Delling, and S. Muller. 1995. Fertilizer experiments with N15 labeled nitrogen in perennial seed grasses for studying N-uptake in varying soil depths. p. 226–230. In Proc. Int. Herb. Seed Conf., 3rd, Halle (Saale), Germany. 18–23 June 1995. Martin Luther Universitat Halle-Wittenberg, Halle (Saale), Germany.
- Silvertown, J. 1980. Leaf-canopy-induced seed dormancy in grassland flora. New Phytol. 85:109–118.
- Smith, J.D. 1987. Winter-hardiness and overwintering diseases of amenity turfgrasses with special reference to the Canadian Prairies. Tech. Bull. 1987-12E. Res. Branch, Agric. Canada.
- Sonmor, L.G. 1963. Seasonal consumptive use of water by crops grown in southern Alberta and its relationship to evaporation. Can. J. Soil Sci. 43:287–297.
- Thompson, D.J., and K.W. Clarke. 1989. Influence of nitrogen fertilization and mechanical stubble removal on seed production of Kentucky bluegrass in Manitoba. Can. J. Plant Sci. 69:939–943.
- Thompson, D.J., and K.W. Clarke. 1993. Effects of clipping and nitrogen fertilization on tiller development and flowering in Kentucky bluegrass. Can. J. Plant Sci. 73:569–575.
This article has been cited by other articles:
|
|
|
|
 
H. A. Loeppky and B. E. Coulman
Crop Residue Removal and Nitrogen Fertilization Affects Seed Production in Meadow Bromegrass
Agron. J.,
May 1, 2002;
94(3):
450 - 454.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
