| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
a Dep. of Plant Sciences, North Dakota State Univ., Fargo, ND 58105
b NDSU Carrington REC, 663 Highway 281 North, Carrington, ND 58421
* Corresponding author (Mohamed.Mergoum{at}ndsu.edu).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: FHB, Fusarium head blight • HRSW, hard red spring wheat • MSGY, main stem grain yield • ND, North Dakota • NDSU, North Dakota State University • NUE, nitrogen use efficiency • RCBD, randomized complete block design • SR, seeding rate • TKW, thousand kernel weight
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication March 28, 2007.
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Tillering plays an important role in HRSW development in the northern Great Plains. Goos and Johnson (2001) found that even though many tillers are initiated, 95 to 100% of the grain yield of HRSW usually comes from the main stem, T1, and T2 tillers in North Dakota. A similar study on spring wheat by Hucl and Baker in Saskatoon, SK (1989) found that, on average, the main stem, T1, and T2 spikes accounted for 67% of the final grain yield. McMaster et al. (1994) reported primary main stem and tiller spikes contributed 83 to 92% of the grain yield depending on timing of irrigation treatments.
Genotypic variances for tiller formation exist among cultivars due to inherent genetic factors (Hucl and Baker, 1989). As a result, cultivars vary significantly in the number of tillers produced with many producing only one or two tillers. Significant differences in percent tiller emergence and survival have been reported in HRSW wheat cultivars (Hucl and Baker, 1989). Low tillering cultivars may respond differently to varying N management and seeding rates compared with cultivars that typically produce high tiller numbers.
Depending on the environment and cultivar, seeding rate can impact wheat tillering and grain yield. Higher seeding rates generally reduce tiller production resulting in more main stem spikes m–2 (Coventry et al., 1993; Hanson, 2001; Staggenborg et al., 2003), which can be favorable for cultivars that tend to produce fewer tillers. Wiersma (2002), however, emphasized that cultivars can differ genetically for yield components and should be tested at different sowing rates to determine their optimum seeding rates. Additionally, optimum seeding rates for HRSW to achieve maximum grain yield vary due to differences in environment and tillering capacity across growing regions of North America. A study conducted on spring wheat in northern North Dakota at Langdon (Hanson, 2001) suggested a minimum plant stand of 2.8 million plants ha–1 (approximately 125 kg ha–1 seeding rate) was needed to attain optimum grain yield. This same study showed that maximum grain yield was achieved at a plant stand of 3.6 million plants ha–1 (approximately 163 kg ha–1 seeding rate). Staggenborg et al. (2003) found the optimum seeding rate of winter wheat to be highly dependent on environmental conditions during the growing season.
Hard red spring wheat is a cool-season crop that is most productive when planted early since cool weather from emergence to the early reproductive stage favors tilling and the development of larger spikes (Wiersma and Ransom, 2005). Therefore, an early seeding date is important in maximizing tiller production and HRSW grain yield. Black and Siddoway (1977) found tiller formation was reduced in HRSW by delayed seeding dates. Similarly, Hucl and Baker (1989) found earlier seeding dates promote more tillers, but more rapid emergence occurred with delayed seeding. Black and Siddoway (1977) also reported significant reductions in spikes m–2 in HRSW wheat when planting was delayed. Hucl and Baker (1989) found earlier seeding resulted in significantly more spikes m–2.
Nitrogen represents one of the most important and expensive inputs in wheat production. The response of spring wheat to N fertilizer varies with rate and timing of application relative to plant development (Mossedaq and Smith, 1994). Their studies have found that N demand increases sharply just before stem elongation (Feekes stage 5), the most rapid phase of crop growth. Foliar applications of N have often been used as a way to increase N use efficiency (NUE) (Sowers et al., 1994). Applying N when it most likely increases crop yield can improve wheat NUE.
Spring wheat tiller production can be influenced by N level and timing of N application. Strong (1986) reported that tillering was enhanced when N was applied before planting or during the tillering process. Weisz et al. (2001) found tiller density increased when N was applied at Zadok's growth stage (GS) 25 (Zadoks et al., 1974) while Oscarson et al. (1995) reported the effect of additional nitrate on the main shoot and tiller number varied, depending on the time of application.
In higher rainfall regions, spring wheat production and quality is often constrained by Fusarium head blight (FHB) [caused by Fusarium graminearum Schwabe (telomorph Gibberella zeae (Schwein.) Petch)]. Some fungicides are partially effective in reducing FHB when applied at anthesis (Jones, 2000). Managing for uniformity in the emergence of spikes is therefore an important objective in highly productive regions where FHB is problematic. Since tillers tend to reach anthesis slightly later than the main stem, reduced tillering may improve the level of FHB control that can be achieved with a single application of fungicides though definitive research on this hypothesis is lacking (Ransom et al., 2007).
The objectives of this research were to: (i) determine the effect of cultivar, seeding rate, N level, and N timing on tiller development, (ii) to understand the relative contribution of main stem and tiller spikes to final grain yield under these diverse management practices, and (iii) to determine if tiller production could be reduced while improving or maintaining yield.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Two cultivars, Alsen and Granite, were used in this study. Cultivar selection was based on tillering and yield potential, disease resistance, straw strength, and plant height. Alsen, a semidwarf cultivar, has average grain yield, medium early maturity, medium tillering potential, strong straw, and moderately resistant to leaf rust and FHB. Alsen has been the most widely grown cultivar in North Dakota since 2003 because of its relatively good resistance to FHB. Granite exhibits high grain yield, medium late maturity, low tillering potential, very strong straw, moderately resistant to leaf rust, and moderately susceptible to FHB.
The experimental design was a RCBD with a split-split plot arrangement with N level as the main plot, N application timing as the subplot, and a factorial combination of cultivar and seeding rate as the sub-subplot. Treatments were replicated four times. Each treatment consisted of a seven-row plot 1.22 m wide and 2.44 m long. Each plot was planted with an Almaco (Seedburo Equip. Co., Chicago, IL) cone-type drill with 15-cm row spacing. Border plots of equal size were planted on the edges of blocks to minimize any border effect.
Two seeding rates, based on a pure live seed basis, consisting of 2.9 million (low seeding rate) and 4.2 million (high seeding rate) seeds ha–1, were used in this experiment. The low seeding rate is based on the current NDSU recommended seeding rate for spring wheat while the high seeding rate is recommended by intensive wheat-management groups to promote less tillering and more main stem heads.
Earliest possible planting dates were used at each location. In 2003, the planting dates were 29 April at Casselton and 24 April at Carrington. In 2004, the planting dates were 17 and 27 April at Casselton and Carrington, respectively. In 2005, the Casselton plot was planted on the 28 April.
Two N levels included in the study were determined based on the N requirement of an average and highest likely attainable yield for each location. At Casselton, N rates were 140 and 224 kg ha–1, targeting yields of 3362 and 5380 kg ha–1, respectively. At the irrigated site in Carrington, 168 and 280 kg ha–1 N were applied targeting yields of 4034 and 6725 kg ha–1, respectively. The actual amount of N applied at each experiment was adjusted for NO3–N levels found in the top 60 cm of soil. Residual NO3–N levels were: 79, 56, 73, 89, and 78 kg ha–1 at Casselton (2003), Carrington (2003), Casselton (2004), Carrington (2004), and Casselton (2005), respectively.
Three N timings were used in this study, preplant, two-way split, and three-way split. For the preplant N treatment, all N was applied as dry urea (46–0–0) and incorporated into the soil with a field cultivator before planting. In the two-way split, N was applied at preplant and at the five-leaf stage (GS 15) at a 1:1 ratio. In the three-way split, N was applied at preplant, at the five-leaf stage (GS 15), and at postanthesis (GS 69) at 1:1:1 ratios. Post-planting N applications were made with urea ammonium nitrate solution (280 g N kg–1). Streamer bars (Amity Technology, Fargo, ND) were used to minimize leaf burn. A flat-fan nozzle was used for the GS 69 application to ensure thorough spike and flag-leaf coverage for maximum absorption.
Weeds were initially controlled with an application of bromoxynil octanoate ester (3, 5-dibromo-4-hydrozybenzonitrile) and MCPA isooctyl ester [(4-chloro-2-methylphenoxy) at 0.28 kg a.i. ha–1 and fenoxaprop-p plus mefenpyr (ethyl 2-[4-[(6-chloro-2-benzoxazoly) oxy] phenoxy] propanoate at a rate of 0.067 kg a.i. ha–1 at GS 15. Additional hand weeding at later stages was also used as needed. To control leaf spotting diseases, 0.126 kg a.i. ha–1 propiconazole {1-[[2-(2,4-dichlorophenyl)–4–propyl-1,3-dioxolan-2-yl] methyl]-1H-1,2,4-triazole} was applied at the early boot stage (GS 39). To control FHB, 0.063 kg a.i. ha–1 of tebuconazole [a-(2-(4-chlorophenyl) ethyl)-a–(1–1-dimethylethyl)–1H-1,2,4-triazole-1–1-ethanol] was made at the beginning of flowering (GS 61) and a second application 7 d later.
Agronomic data were collected on stand density, tiller density, spike density, tillers per plant, tiller mortality, grain yield, 1000-kernel weight (TKW), kernels per spike, and yield of the main spike and primary tiller (T1, T2, and T3) spikes. Stand density was measured at GS 12 by averaging the total number of plants in two random 61-cm-long row sections. The number of total stems at GS 37 and spike number at GS 75 were taken on the same row sections used for plant stand. Tiller density was determined by subtracting the stand count from the total stem number at GS 37. Tillers per plant were determined by dividing the tiller density by the stand density. Spike number was determined by averaging the total number of fully developed spikes at GS 75. Tiller mortality was determined by subtracting the spike density from the tiller density and dividing by the tiller density.
Before harvest, a random 30-cm section of one row was sampled in two replicates to determine grain yield contribution among the main stem spikes vs. T1, T2, and T3 tiller spikes. Culms were identified according to the morphological naming scheme of Klepper et al. (1983). The main stem spike was identified first on the primary culm. The T1, T2, and T3 tiller spikes were then identified as the primary tiller emerging from the axil of the first, second, and third true leaf, respectively, on the main stem. Each 30-cm section of row was identified, counted, and labeled for the main stem, T1, T2, and T3 spikes. Spikes were then threshed on an Agriculex SPT-1 belt threshing machine (Agriculix Inc., Guelph, ON). Seeds were collected, weighed (g), and totaled for each plot. The contribution of individual spikes to grain yield (%) was determined by dividing the weight of the main, T1, T2, and T3 spikes by total weight of all spikes combined for each plot.
Following harvest, grain samples were dried using forced natural air to an equilibrium moisture content of approximately 12% as determined by testing several samples using a Motomco moisture meter (Motomco Inc., Paterson, NJ). Grain samples were cleaned on a clipper grain cleaner and weighed to determine grain yield (kg ha–1). Thousand kernel weights (g) for each plot were determined by weighing 250 kernels obtained from five spikes selected at random from each plot in two replicates. Kernels per spike were determined by averaging the number of kernels obtained from five spikes selected at random from each plot in two replicates. For both TKW and kernels per spike, the kernels were counted using an electronic seed counter (Seedburo Equipment Co., Chicago, IL) after being threshed on an Agriculex SPT-1 belt threshing machine.
Data analysis was performed using SAS (SAS Institute, 2004). Data from individual environments (location and year) were analyzed separately using ANOVA. A combined ANOVA was conducted across environments using data from all five environments. In the combined analysis, environment was considered a random effect, while cultivar, seeding rate, N level, and N timing were considered fixed effects. Means were separated using an F-protected LSD (P = 0.05) as described by Steel and Torrie (1997).
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
|
The SR x C interaction for tiller density was also significant (Table 2) but due to differences in magnitude (data not shown). Increasing the N level reduced tiller number by 35 tillers m–2 (Table 3). Increasing the seeding rate decreased tiller density by 43 tillers m–2 (Table 4). This response may have resulted from competition due to the increased seeding rate, which may have reduced the survival of tillers. Similar results on tiller density were reported by Hanson (2001). Tiller density was also reduced with the three-split N application compared with the preplant and two-split applications perhaps due to increased N availability with the preplant and two-split N treatment during the tillering phase of development (Table 4). Differences in tiller density based on timing of N application were also reported by Oscarson et al. (1995) in which the response to additional nitrate varied depending on the time of N application. Other studies (Weisz et al., 2001; Strong, 1986) have also focused on the timing of N application rather than N rate. Weisz et al. (2001) found that an N application at GS 25 had a significant effect on tiller production. When no N was applied at GS 25, there was a significant decrease in tiller density by GS 30. Weisz et al. (2001) also found tiller density increased as N rate was increased up to a maximum tiller density at 101 kg ha–1. These results differ from this study in which maximum tiller density was achieved at the lowest N rate. Additionally, no significant increase in tiller density occurred with the two-split N application compared with the preplant N application.
Tillers per plant was significantly influenced by the C x E and seeding rate x environment (SR x E) interactions and N timing, and seeding rate main effects. The significant C x E interaction was primarily due to Alsen producing an average of 1.4 tillers per plant compared to 0.8 tillers per plant for Granite at Casselton (2003) (Table 3). At Carrington (2003), Granite produced an average of 1.2 tillers per plant while Alsen produced 1.0 tillers per plant. At both locations in 2004 and 2005, no significant difference in the number of tillers per plant between cultivars occurred (data not shown). The SR x E interaction was significant for tillers per plant, but due to differences in magnitude only. The main effect of increased seeding rate reduced tillers per plant by 36% (Table 4) due to increased competition. Similar results of increased seeding rate on tillers per plant were reported by Hanson (2001). Among N timing treatments, the preplant and two-split applications were equal, and produced an average of 0.3 more tillers per plant compared to the three-split N application (Table 4). The increase in tillers per plant with the two earlier N applications may have resulted from additional N available to the plant during the tillering phase (GS 20–25) of plant development allowing for additional tiller development, while the three-split N application would be of no benefit since it was applied after the tillering process was already completed (post-anthesis).
Spike density was influenced by N level, N timing, and seeding rate main effects. Increasing seeding rate significantly increased spike density by 37 spikes m–2 when averaged across cultivars, N levels, and N timing treatments (Table 4). Similar results on the effect of increased seeding rate on spike density were reported by Coventry et al. (1993) and Hanson (2001). Spike density was increased by 58 spikes m–2 when N level was increased (Table 4). Another study by Mossedaq and Smith (1994) found similar results of increasing N level on spike density. Spike density was significantly reduced by 44 spikes m–2 with the three-split N timing compared to the preplant treatment (Table 4). The reduction in spike density was likely associated with the corresponding reduction in tiller number. This response may have been due to the delayed three-split N application not benefiting these yield components since the N application occurred after these components were already determined. From these data, the earlier N applications (preplant and two-split) resulted in an increase in tiller number and corresponding increase in spike density.
Tiller mortality was significantly influenced by the N timing x cultivar (T x C) interaction. Across N-timing treatments, Alsen's tiller mortality remained unchanged at about 3.5%. Granite's tiller mortality was significantly reduced with the three-split N application. Tiller mortality for Granite was 4.6, 3.0, and 2.3% for the preplant, two-split, and three-split N timing treatments, respectively (data not shown).
The significant C x E interaction for TKW was due to differences in magnitude due to Alsen producing a greater TKW in four out of five environments. Alsen's TKW was 2.8 to 5.0 g greater than Granite at Casselton (2004 and 2005) and Carrington (2003 and 2004) and equal to Granite at Casselton 2003 (Table 3). The significant N level x N timing (N x T) interaction for TKW was also due to differences in magnitude (data not shown). Nitrogen timing main effects also significantly influenced TKW. Thousand kernel weight for the preplant and two-split N-timing treatments were equal at 35.3 g, while the three-split treatment produced a TKW of 36.2 g across five environments (Table 4).
Nitrogen timing x environment (T x E) and C x E interactions, and cultivar and seeding rate main effects significantly influenced kernels per spike. The T x E interaction significantly influenced kernels per spike due to a decrease in the number of kernels with the three-split N timing at Carrington 2004. Kernels per spike failed to differ significantly for the preplant and two-split N timing treatments with 33.5 and 32.5 kernels per spike, respectively, while a significant reduction occurred for the three-split treatment with 26.9 kernels per spike at Carrington 2004 (Table 5 ). The Casselton 2003 through 2005 and Carrington 2003 locations showed no significant difference among N timing treatments for kernels per spike. The C x E interaction for kernels per spike was due to differences in magnitude with Granite consistently producing more kernels per spike than Alsen across all five environments (Table 3). Granite produced from 3 to 13 more kernels per spike than Alsen across environments. Seeding rate main effects significantly influenced kernels per spike (Table 2). As seeding rate was increased, kernels per spike decreased from 31.1 to 29.4 kernels per spike across environments (Table 3).
|
The C x E interaction for grain yield was significant (Table 2). At Casselton (2003), Alsen produced 350 kg ha–1 higher grain yield compared to Granite under dryland conditions, while at Carrington (2003), Granite produced 463 kg ha–1 higher grain yield compared to Alsen (Table 3). In 2004, Granite produced 215 and 967 kg ha–1 greater grain yield than Alsen at both Casselton and Carrington locations, respectively. At Casselton (2005), Granite again produced a significantly greater grain yield of 4232 kg ha–1 compared to Alsen at 3782 kg ha–1. Granite's increase in grain yield compared to Alsen in four out of the five environments may have been associated with Granite's ability to withstand lodging under heavier rainfall or irrigated conditions.
Tiller Contribution to Grain Yield
The level of significance of main effects and interactions to main and tiller spike contributions to grain yield from the ANOVA are presented in Table 2. The C x E and SR x T interactions and environment and seeding rate main effects significantly influenced the main spike contribution to grain yield. No significant differences among cultivar, N level, or N timing main effects in the contribution of the main spike to grain yield were found.
The C x E interaction influenced main spike contribution to grain yield. At Casselton 2003, Alsen and Granite's main spike contributed 27.9 and 23.4% to grain yield, respectively (Table 3). However, at Carrington 2004, Alsen and Granite's main spike contributed 52.5 and 57.4% to grain yield, respectively. The difference between cultivars at Casselton 2003 in percent main spike contribution to grain yield could be attributed to the drier growing conditions which may have prevented Alsen from producing more tillers resulting in a greater contribution of the main spike to grain yield. Granite, which tends to produce fewer tillers, may have performed better under irrigated conditions at Carrington 2004 resulting in a greater main spike contribution to grain yield.
At the lowest seeding rate, the three-split N timing increased the main spike contribution to grain yield by 3.6 and 2.4% compared to the two-split and preplant, respectively (Table 6 ). At the high seeding rate, no significant difference among the N timing treatments for main spike contribution to grain yield occurred possibly due to increased competition for nutrients and light. The presence of more N early in the growth cycle promoted tillering when plant spacing was low enough to permit additional tiller production.
|
The T2 spike contribution to grain yield was significantly influenced by environment and seeding rate main effects. As may be expected, increasing the seeding rate reduced the T2 spike contribution to grain yield by 2% (Table 7 ). The N x E interaction was due to differences in magnitude, with the T2 contribution being consistently higher at the highest N rate across all environments (data not shown). The T2 spikes consistently contributed only about 11% to the overall yield regardless treatment factor. The low contribution of the T2 spike to yield was due to both the absence of these spikes, as less than half of the plants had T2 spikes at harvest (Table 7) as well as to their reduced size.
|
Overall, in this study, the main and T1 spikes contributed to 86 and 87% of the final grain yield of Alsen and Granite, respectively. Previous work conducted in North Dakota on HRSW tiller contribution to grain yield by Goos and Johnson (2001) found that wheat grain yield in North Dakota comes almost entirely from the main, T1, and T2 spikes. Similarly, other studies with HRSW (Hucl and Baker, 1989) found that, on average, 67% of the final grain yield came from the main, T1, and T2 spikes. The data in this study demonstrated that the main and primary tiller spikes (T1) were most important in contributing to overall grain yield with the T2 and T3 tillers contributing significantly less. Additionally, the relative contribution of the main and tiller spikes can be changed by altering the seeding rate.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
