Wheat Variety Response to Seeding Rate in Kansas During the 2015–2016 and 2016– 2017 Growing Seasons

This report is brought to you for free and open access by New Prairie Press. It has been accepted for inclusion in Kansas Agricultural Experiment Station Research Reports by an authorized administrator of New Prairie Press. Copyright 2018 the Author(s).


Introduction
Plant density is among the major factors determining the crop's ability to capture resources such as water, nutrients, and solar radiation (Satorre and Slafer, 1999). The response of wheat to plant density is largely determined by competition for resources with neighboring plants, and increased competition can result in reduced survival, dry matter production, and grain yield of individual wheat plants (Satorre, 1988). Wheat plants subjected to high density generally have fewer tillers and grains than widely spaced plants (Rana et al., 1995). On the other hand, too widely spaced plants can result in few plants per unit area and consequently less grains per unit area, explaining the typical parabolic response of grain yield to plant density (Holliday, 1960). Consequently, appropriate management of population density may allow maximum yields per unit area to be achieved (Satorre and Slafer, 1999). Given the difference in wheat varieties regarding their ability to tiller as well as their response to intra-canopy competition for resources, it is possible that different varieties require different seeding densities to maximize yield. Therefore, the main objective of this project was to better understand the response of different wheat varieties to seeding rate.

Procedures
One experiment was conducted at four site-years: at the South Central Experiment Field near Hutchinson, KS, during 2015KS, during -2016KS, during and 2016KS, during -2017 at the Agronomy North Farm in Manhattan, KS, during 2015KS, during -2016andat the North Central Experiment Field in Belleville, KS, during 2016-2017. Trials were established in a randomized complete block design with four replications. Seven varieties (i.e. Everest, KanMark, 1863, Larry, Zenda, Tatanka, and Joe during 2015-2016andKanMark, Larry, Zenda, Tatanka, Joe, Bob Dole, andAG Icon in 2016-2017) and five seeding rates (0.6, 0.95, 1.3, 1.65, and 2 million seeds/a) were tested, for a total of 35 treatments and 140 plots per location. Plots were 7 rows wide at a 7.5-in. row spacing in Manhattan and at both locations during the 2016-2017 growing season, and 6 rows wide at a 10-in. row spacing in Hutchinson. The harvestable portion of the plots was approximately 20-ft long at all locations.
Management practices adopted at all locations are described in Table 1 and initial soil fertility is listed in Table 2. Nitrogen (N) fertilization at all locations was performed with a yield goal of approximately 70 bu/a. Weeds and foliar diseases were controlled at both locations. Agronomic measurements included stand count approximately 3-4 weeks after planting, percent canopy cover measured several times during the growing season using digital imagery, and a 1-meter row subsample clipped from each plot at harvest time for biomass, harvest index, head count, average grain weight, and head size. The latter samples were still being processed at the time this report was prepared, therefore, results are not shown in the current report. Plots were harvested using a small plot combine at all locations, and grain yield was adjusted to a 13% moisture basis.

Growing Season Weather
The weather during the 2015-2016 growing season was characterized by a warm and moist fall, followed by a dry and mild winter and a cool and moist spring (Table  3). Meanwhile, the 2016-2017 growing season started with a drier fall with similar temperature totals, received earlier moisture during the winter, and had a similar spring to that observed during the previous season, with plenty of precipitation and belowaverage temperatures (Table 3). Growing season precipitation total was 20.5 in. in Hutchinson and24.4 in. in Manhattan (2015-2016), and 18.2 in. in Hutchinson and14.8 in. in Belleville (2016-2017). Despite the high precipitation totals, cumulative solar radiation during the growing season was well above 3,000 MJ m -2 at all studied site-years, indicating that solar radiation should not have been a yield-limiting factor in this study.

Stand Establishment
The trials were sown into adequate moisture at all locations, which ensured good germination and stand establishment. Average percent establishment (final stand over targeted seeding rate) was 72% in 2015-2016 and 92% in 2016-2017. At all site-years, increasing seeding rate increased the final stand count for all varieties at all locations ( Figure 1).

Wheat Grain Yield: Individual Site-Year Analysis
There was a great difference in yield potential among study-locations, with average yield across all varieties and plant population densities ranging from 44 bu/a in Manhattan 2015-2016, 78 bu/a at both Hutchinson 2015-2016 and Belleville 2016-2017, and 101 bu/a in Hutchinson 2016-2017 ( Figure 2). Yields were normally distributed across all locations. At all individual studied locations, grain yield was significantly affected by variety and by planting density, but there was no significant interaction (Table 4). In other words, there were grain yield differences among varieties and among population densities; however, the different varietal responses to planting density were not captured in each individual site-year analysis (all varieties responded similarly to the change in population density in each individual location). At all locations and years, wheat grain yield response averaged across varieties was well represented by an exponential rise to the maximum on a non-linear regression model, with wheat grain yields reaching 95% of the asymptotic maximum at approximately 890,000-911,000 emerged plants per acre in three out of four sites ( Figure 3). The only exception was Hutchinson during 2015-2016, when grain yields maximized at 530,000 plants per acre. The lowest population density treatment at each location, which ranged from 445,000 to 721,000 plants per acre depending on site-year, resulted in grain yields statistically similar to the very next plant population density at all site-years, but had lower yields than the following greater population density treatments (greater than approximately 850,000 to 1,000,000 plants per acre, Figure 3).

Wheat Grain Yield: Analysis Pooled Across Site Years
The pooled analysis of variance was first performed over the entire dataset using raw yield data. Subsequently, due to the differences in yield environment among the four site-years in this study (Figure 2), the analysis was performed using relative yields. Relative yields were calculated for each variety at each site-year using the highest yielding plot for a particular variety as the denominator for all plots for that same variety. Wheat varieties behaved differently at each location and year, but some trends were observed. Grain yield averaged across seeding rates for each variety is shown in The initial analysis using the raw yield data allowed us to screen for varieties more responsive to plant population (i.e. varieties that showed large yield increases at higher stands), average responsive varieties, and less responsive varieties (varieties that tended to maximize yields at very low seeding rates). Among varieties that maximized yields at low seeding rates were Tatanka, Everest, and 1863, all of which maximized yields between 650,000 and 695,000 plants per acre ( Figure 5). It is important to highlight that Everest and 1863 were only tested during 2015-2016 and thus reflect only one year's data, which gives us less confidence in the results. Tatanka has now a total of two years of data, providing greater strength to assume its good performance under low population densities. The majority of the varieties belonged to the average response group, including Joe, KanMark, Zenda, and Bob Dole -the latter only evaluated in one year of the experiment. This group maximized yields between 785,000 and 900,000 plants per acre ( Figure 5). Varieties that required more plants to maximize yields included Larry and AG Icon (single year of data for the latter one), which required 1,080,000 to 1,060,000 plants per acre to maximize yields ( Figure 5).
The previous results were later confirmed by the subsequent analysis, which discriminated among varieties but evaluated relative rather than raw grain yield. In the relative grain yield analysis pooled across site-years, Joe, KanMark, and Zenda again maximized yields between 775,000 and 870,000 plants per acre, which reflects the average response group (Figure 6); Larry and AG Icon maximized yields at populations beyond 950,000 plants per acre (more responsive varieties; Figures 6 and 7); and Tatanka, Bob Dole, Everest, and 1863 maximized yields at populations less than 690,000 plants per acre (less responsive varieties; Figures 6 and 7). Everest and 1863 showed no significant response to plant densities (Figure 7). Bob Dole was the only variety that showed discrepant results between the relative yield and raw yield data analyses, as it was categorized as an average responsive variety using the raw data and a less responsive variety using the relative yield data. Results for Everest, 1863, Bob Dole, and AG Icon should be interpreted with more caution than the remaining ones because they only reflected one year's data, and more tests are needed to increase the power of the analysis.

Preliminary Conclusions
With four site-years of data, we start gathering firm conclusions about each variety's response to plant population. Zenda, KanMark, and Joe seem to have an intermediate response to seeding rate and maximize yields around 800,000 to 900,000 plants per acre. Tatanka seems to be less responsive to plant population, maximizing yields with populations as low as 565,000 to 660,000 plants per acre. Larry has shown greater response to plant population, and yield was only maximized at populations above 1,060,000 plants per acre. While preliminary data suggest Everest and 1863 are not responsive to plant population, Bob Dole is intermediate, and AG Icon is more responsive, the limited number of observations (two site-years of data only) limit the power of this analysis and the breadth of these conclusions, not allowing for broader implications from the data. This study is currently being conducted at five locations during the 2017-2018 growing season so that more definite recommendations can be drawn for each variety.