memb1ers

Tiller mortality and its relationship to grain yield in spring wheat

A primary determinant of grain yield in barley (Hordeum vulgare L. emm. Lam) is the number of ear-bearing tillers per plant at harvest, which depends both on the production of tillers and on their subsequent survival to form ears. This three-year field study compares tiller production and survival in relation to final grain yield in three types of barley: 2-rowed winter (2rw), 6-rowed winter (6rw) and 2-rowed spring (2rs), grown in two contrasting environments. These three types differed significantly in shoot and ear number, the winter barleys showing higher tiller production, with the maximum number of tillers ranging from 798 to 2315 m−2 in 2rw, 711 to 1527 in 6rw and 605 to 1190 in 2rs. Grain yield across environments and years was strongly correlated () with the number of ears at harvest. The maximum number of shoots produced by each type of barley was inversely related to the mean temperature during the tillering phase. Tiller mortality was inversely related to the maximum shoot production, being significantly lower in barleys with less tillering capacity, i.e. the spring type (with average values of 34.3% and 42.7% in the two environments). The highest tiller mortality occurred before anthesis and, to a lesser extent, from anthesis to maturity. These data support the hypothesis that the principal cause for tiller mortality in barley grown under Mediterranean conditions is the competition between tillers for a limited supply of resources.



Spikeless tillers of wheat (Triticum aestivum L.) affect grain yield because of less than optimum effective plant population. This study was conducted to examine the genetic variability for tiller mortality, and its relationship to grain yield in diverse wheat lines. Twenty lines were evaluated in replicated field tests in 4 years at Rampur, Nepal. The characters investigated were maximum number of tiller produced, the number of reproductive tillers, tiller mortality, and grain yield. The lines differed significantly for all characters. The tiller mortality ranged from 7 to 30%. There were substantial effects of environment on all four characters. The entry-by-year interactions were significant for all traits, primarily because of changes in the relative genotypic differences for these traits in the four years. However, certain lines consistently ranked low or high for tiller mortality. There was a significant negative correlation between front tine tiller and grain yield in 3 out of 4 years. There was a positive correlation of highest tiller number with reproductive tiller number and with tiller mortality. Grain yield showed a nonsignificant positive correlation with maximum tiller number. The reproductive tiller number was positively correlated with grain yield. Results of this study indicate that spikeless tillers contribute negatively to grain yield and that genetic variation exists for tiller mortality in spring wheat.



Vegetative growth in the form of tillers is crucial to final yield in winter wheat (Triticum aestivum L.). To understand the impact management practices have on tiller initiation, a study was conducted using two seeding rates (1.9 × 106 vs. 6.8 × 106 ha−1) and two N timing applications (single vs. split). Tillers initiated in the fall made up the majority of spikes compared to tillers initiated from 1 January to the start of jointing (GS 30). Tillers initiated in March at either seeding rate produced very few kernels spike–1, low kernel weight, and contributed little to yield. At the high seeding rate, tillers initiated prior to 1 January were responsible for more than 87% of the grain yield. Tillers produced in January– February produced 5 to 11% of the final yield, while tillers produced in March contributed less than 2%. In contrast, at the low seeding rate tillers produced in January–February made up 20 to almost 60% of the final yield. Overall, this study shows the timing and rate of leaf initiation impacts yield and yield components. Earlier tillers have an advantage in that they have shorter periods of leaf development that result in more leaf area which in turn supports more kernel spike–1 and heavier kernels, thus more grain weight per spike. Timing of N (single vs. split) application resulted in no significant impact on tiller development, spike number, kernel number, kernel weight, or grain yield.



The number of spikes ha–1 is a critical yield component of wheat yield. Two factors contribute to the total number of spikes ha–1 at harvest, number of mainstem (MS) spikes and number of tillers plant–1. The number of tillers produced per plant is controlled by the environment during the period of tiller development from three-leaf stage to jointing (GS13–GS30) (Klepper et al., 1982) and the amount of tiller mortality that occurs from jointing to anthesis (GS30–GS69) (Jewiss, 1972; Rawson, 1971). Recent research has shown that the timing of tiller initiation and management factors such as seeding rate influence the rate of leaf development on each tiller which, in turn, influences tiller size and mortality (Tilley et al., 2015). The timing of tiller initiation and management factors such as planting date (Oakes et al., 2016) that promote leaf development could also influence other yield components such as kernels spike–1 and kernel weight. An understanding of when the most spikes are formed and the management factors that promote tiller formation during this critical period would help growers improve wheat yield.



Tillers can be formed at multiple nodes on the MS, and secondary and tertiary tillers can form from nodes on the tillers themselves (Klepper et al., 1982; Evers and Vos, 2013). Under glasshouse conditions Klepper et al. (1982) found that once a tiller is initiated, leaf development on the tiller proceeded at the same rate as leaf development on the MS. However, subsequent research has found that leaf development on each tiller proceeds at a slower rate than that on the MS or even on preceding tillers (Tilley et al., 2015). This indicates that tillers initiated first will always have an advantage in growth and development compared to those initiated later. This advantage will increase as time passes resulting in more leaf area. It is likely that tillers with more leaf area will produce more kernels, heavier kernels, and will be less likely to be lost to tiller mortality.



Timing of tiller initiation can also influence tiller mortality. Charles-Edwards (1984) concluded that self-thinning within plant communities is largely due to the lack of assimilate needed to continue growth and development within the individual stem which, in turn, can lead to a decrease in plant weight and eventually a decrease in plant yield. Some works have explored the purpose of rear tine tiller and the effects it may have on the plant as a whole and concluded tillers that abort may have benefited the plant due to assimilate and nutrient accumulation (Lupton and Pinthus, 1969; Palfi and Dezsi, 1960). However, Langer and Dougherty (1976) concluded that dead tillers had a negative effect on grain yield due to competition for assimilates and nutrients (Sharma, 1995).



Management practices such seeding rate and N application timing can influence the timing and rate of leaf and tiller development (Bauer et al., 1984; Tilley et al., 2015) and grain yield. Tompkins et al. (1991) concluded that grain yields will decline as seeding rates decline. This in part is due to a decrease in spikes. However, it was determined that grain yield can decrease at high seeding rate (HSR) (Gooding et al., 2002) due to a decrease in kernels spike–1 and a decrease in kernel weight (Puckridge and Donald, 1967; Tompkins et al., 1991). Tilley et al. (2015) found that seeding rates influenced the rate of leaf development. Phyllochron intervals (PI) were shorter for each tiller at a low seed rate (LSR) compared to the same tillers at a HSR. This resulted in more leaves on each tiller, more tillers produced and fewer tillers lost to tiller mortality.



Nitrogen is recognized as a vital nutrient needed for growth and development (Miller, 1939; Wilhelm et al., 2002). Nitrogen application timing recommendations for winter wheat in North Carolina (NC) are based on the tiller density (Weisz et al., 2001, 2011). Winter split applications are encouraged if tiller density <550 m–2. Otherwise the standard NC recommendation is to apply N at GS 30, the time when the wheat stem begins to elongate. Maidl et al. (1998) confirms that early N application increased plant density and concluded that N fertilizer treatment applied during stem elongation not only reduced tiller mortality but also led to high grain yield in both MS and tillers.



To understand the impact of the timing of tiller initiation and management practices on kernel development and yield, a method of counting and marking leaves and tillers was created to monitor tiller growth and decline. This monitoring of individual tillers resulted in the ability to measure the number of heads, kernel number and kernel weight each tiller produced, and its contribution to final yield. The objectives of this study were to: (i) measure yield and yield components of tillers initiated at different periods during the growth of wheat and how tillers initiated at different periods contribute to overall grain yield, and (ii) determine the impact of seeding rate and timing of N applications on the productivity and sustainability of tillers initiated at different periods during the growth cycle of wheat.



MATERIALS AND METHODS

Field Experiment

Field experiments were conducted at two sites in eastern NC and one site in western NC. At the Tidewater Research Station (TRS) in Plymouth, NC, experiments were conducted in 2009, 2010, and 2011. On a private farm in Beaufort County (BC) experiments were conducted in 2009 and 2010. On the third site in western NC (Piedmont Research Station [PRS] in Salisbury, NC) a single trial was conducted in 2011. The soil at TRS was a Cape Fear loam (clayey, mixed, thermic Typic Umbraqult) soil. At the BC site in 2009 and 2010 the experiment was conducted on a Cape Fear fine sandy loam (clayey, mixed, thermic Typic Umbraqult). The 2011 experiment at PRS was conducted on a Mecklenburg clay loam (fine, mixed, thermic Ultic Hapludult). In 2009, plots were planted on 3 November at TRS and 4 November in BC. In 2010, plots were planted on 10 November at TRS and 11 November in BC. In 2011, plots were planted on 10 November at TRS and 15 November at PRS.



At each site, Pioneer 26R12, a high yielding wheat variety in NC, was planted in 16.9-cm rows into a conventional tilled field following corn. The experimental design at all sites was a split plot design with main plots consisting of two seeding rates, 1.9 × 106 and 6.8 × 106 ha–1, and subplots consisting of 134 kg N ha–1 applied either as a single application in March or a split application with half applied in late January or early February and the remaining half applied by late March. In 2009–2010, the first N application was made on 15 February with the second split and single N application made on 22 March. In 2010–2011, the first N application was applied on 4 February while the remaining split and single applications were completed on 18 March. During the 2011–2012 growing season at TRS, the first split application was applied on 19 January with the final split and single N applications applied on 12 March. Applications at the PRS were applied 1 wk later on 26 January and 19 March. All treatments were replicated five times.



Disease pressure was minimum across all three site years and did not reach current threshold recommendations (Weisz et al., 2011). However, weed and insect control practices were applied. In 2009–2010 at TRS, thifensulfuren-methyl/tribenuron-methyl was applied POST at 0.04 kg a.i. ha–1 on 8 Mar. 2010. The BC location received the same application on 9 Mar. 2010. In 2010–2011 at both TRS and BC, thifensulfuren-methyl/tribenuron-methyl was applied POST at 0.04 kg a.i. ha–1 on 14 Mar. 2011. In 2011–2012 at TRS, mesosulfuren-methyl was applied POST at 0.33 kg a.i. ha–1 on 6 Dec. 2011 and thifensulfuren-methyl/tribenuron-methyl applied POST at 0.05 kg a.i. ha–1 on 1 Jan. 2012. At the PRS, chlorsulfuron/metsulfuron-methyl was applied PPE at 0.03 kg a.i. ha–1 on 3 Nov. 2011 and thifensulfuren-methyl/tribenuron-methyl was applied POST at 0.05 kg a.i. ha–1 on 28 Feb. 2012.



Individual plots were 24.4-m long and 1.98-m wide equaling a total of 48.31m2. Each plot was divided into three sections. The first 18.01 m2 section was designated for grain yield and grain sampling. This section of the plot was harvested using a Gleaner K2 combine with a Harvestmaster Graingage (Juniper Systems, Logan, UT) that recorded moisture, grain weight, and test weight. The TRS in 2010–2011 was harvested on 20 June and on 22 June during 2011–2012. Beaufort County in 2010–2011 was harvested on 23 June and PRS was harvested on 29 June during the 2011–2012 season. Grain weight was adjusted to 15.5% moisture before calculating yield.



The second section equaling 9.12 m2 was designated for marked samples. Five plants from each plot were marked and the number of full and partial leaves on each MS and tiller were recorded along with the total number of tillers at current growth stage. This was done once a month from planting to harvest. Throughout the 2009–2010 growing season, observations were made at TRS and BC on 22 December, 28 January, 1 March, 19 March, 7 April, and 26 April. During the 2010–2011 growing season, observations were made on 7 December, 31 January, 4 March, 2 April, and 30 April. During the 2011 growing season, leaf and Garden Tiller and Cultivator counts were recorded on 9 December, 2 January, 11 February, and 3 April at TRS and 15 December, 9 January, 24 February, and 13 April at PRS. Each new and existing tiller was noted using either a black, silver, or red permanent marker to mark leaf number. Black markings represented tillers that were initiated from planting through the end of December. Silver markings represented early winter tillers that developed from the first of January to the beginning of March. Red markings represented late spring tillers produced from March till growth stage GS30. The three colors used to track tillers helped categorize each individual tiller and determined whether or not they initiated in the fall, winter, or spring. Furthermore, tillers were marked on each subsequent leaf to track the number of leaves produced throughout the growing season. Harvest samples were taken in 2010–2011 and 2011–2012 at TRS, BC, and PRS on the same dates that the larger plots were harvested. At harvest, each of these five plants were clipped and placed in individual bags. For each plant, the MS and tillers were separated by color markings (black, silver, red) counted and hand threshed to determine the number of spikes and grain weight spike–1 for each tiller initiation period. The data for all five plants were averaged to represent values for each plot.



The last 21.18 m2 of each plot was reserved for destructive sampling. Method for destructive sampling consisted of a 2-m stick and a garden shovel. A trench, encamping an area of 0.33 m2, was carefully dug around plants to a depth of 15 cm and the plants were then excavated from the destructive sampling area. Samples were taken on 17 June 2010, at TRS and BC. On 15 and 20 June 2011, destructive samples were taken at TRS and BC. Destructive samples were taken at TRS and PRS in 3012 but were destroyed before they could be processed. Leaf counts were taken from each individual stem and recorded. Leaf numbers were determined by counting the nodes on the plant. This was done by splitting the plant at the base and finding the small (0.6–1.25 cm) gap between the compressed nodes and the first separated node. The first separated node was counted as the fifth node (fifth leaf) and subsequent nodes (leaves) were counted in ascending order. Plants were separated into classes corresponding to the periods of tiller initiation (black, silver, and red) based on leaf number and the ratio of stems found in each initiation period in the marked samples. This ratio was determined by counting the number of MS or tillers from each category (black, silver, and red) in the five marked plants described above and dividing that number by the total number of MS or tillers produced in these same plants. Using the ratio of MS or tillers that were initiated from planting to the end of December (Black), the same ratio of plants with the highest leaf numbers in the destructive sample were designated as having been initiated during this period. Plants with the next highest leaf number were considered initiated during the period from 1 January to the end of February; and plants with the fewest leaves were considered initiated after 1 March. Spikes from samples representing each initiation period were hand threshed and grain weight, kernel number and 100 kernel seed weight were measured.



Statistical Procedures

For the marked plant samples the data taken from TRS in 2010–2011 and 2011–2012 at BC in 2010–2011 and PRS in 2011–2012 were analyzed using a repeated measures design with the Proc Mixed procedure in SAS (SAS Institute, Inc., Cary, NC) to determine if there were differences in the number of spikes plant–1 and grain weight spike–1 among site-year, tiller initiation periods (planting to 31 December, 1 January to 28 February, and after 1 March) seeding rate, and N application timing. In all cases, site-year, seeding rate and N timing were treated as fixed effects, while blocks and the interactions with blocks were treated as random. When differences were detected, Fisher’s Protected LSD was used to separate means.



In the destructive sample plots some samples were lost in 2011–2012. Therefore, only samples taken in 2009–2010 and 2010–2011 at TRS and BC were used in the analysis. The Proc Mixed procedure in SAS (SAS Institute, Inc.) was used to determine if there were differences in spikes m–2, kernels spike–1, weight per 100 kernels, and grain yield among site-years, mini power tiller tractor initiation period, seeding rate and N application timing. As with previous analysis, site-year, seeding rate, and N timing were treated as fixed effects; while blocks and the interactions with blocks were treated as random. When differences were detected, Fisher’s Protected LSD was used to separate means.



Grain yield from the large 18.01 m2 section of each plot for the 2010–2011 and 2011–2012 seasons at TRS and BC were analyzed using the Proc Mixed procedure in SAS (SAS Institute, Inc.) to determine if there were differences in grain yield among site-years, seeding rate, and N application timing. These site-years were chosen so that the grain yield from the large samples could be compared with that calculated from the small 2-m samples.