Abstract
Potassium fertilization and harvest timing both influence the stand life and performance of alfalfa (Medicago sativa L.), but many producers do not apply sufficient K fertilizer as recommended by soil test recommendations. The objective of this trial was to determine the impact of reduced rates of K fertilization on cultivar Bulldog 805 alfalfa managed under different harvest regimes on forage yield, stand persistence, and nutritive value when grown in the southern Coastal Plains. Plots were harvested at bud, 10, 30, and 50% bloom growth stages. Potassium was applied three times across the season at total rates of 0, 67, 101, 134, and168 kg K2O ha−1. Aboveground and belowground plant and soil responses were evaluated. Evaluated levels of K were not sufficient to elicit a positive response in aboveground variables, as all treatments exhibited K deficiency as confirmed by tissue analysis at the end of the study. Harvest timing affected many of the aboveground yield components, however no clear trends existed. As expected, material harvested at later growth stages tended to have higher fiber and lower digestibility than alfalfa harvested at less mature stages. Reduced rates of K fertilization helped maintain soil K levels above the final level in the untreated control. Harvesting alfalfa at later (50% bloom) growth stages decreased starch and nonstructural carbohydrate (NSC) content in the roots. To optimize both alfalfa yield and nutritive value, current University recommendations for alfalfa harvest timing and K fertilization should be maintained in the southern Coastal Plains.
Abbreviations
ADF
acid detergent fiber
CP
crude protein
DM
dry matter
ESC
ethanol?soluble carbohydrate
IVTDM
in?vitro true dry matter digestibility
NDF
neutral detergent fiber
NFC
nonfibrous carbohydrates
NIRS
near infrared reflectance spectroscopy
NSC
nonstructural carbohydrates
TDN
total digestible nutrients
1 INTRODUCTION
Alfalfa (Medicago sativa L.) is a high?quality perennial legume that can improve forage–livestock systems with biological N2 fixation, improve forage nutritive values, and increase animal gains. While alfalfa was once the dominant perennial legume species used in the southern United States, the harsh environment (high day and nighttime temperatures, high humidity) and elevated insect pressure soon eliminated many productive alfalfa stands (Lacefield, Ball, Hancock, Andrae, & Smith, 2009). The recent development of alfalfa cultivars well suited for this harsh climate has led to an increased interest in planting alfalfa in the region. It is well known that K fertilization and harvest timing both influence the productivity and longevity of an alfalfa stand (Berg et al., 2009; Gasser et al., 1969; Undersander et al., 2011); however, these management practices need to be examined further in the context of southeastern environments.
Potassium is a critical nutrient for many biological functions and affects overall plant performance (Grewal & Williams, 2002). Deficiencies of K are associated with increased disease, decreased photosynthetic ability, and reduced carbohydrate availability, all of which can negatively influence yield potential in alfalfa through winter kill or stand thinning (Amtmann, Troufflard, & Armengaud, 2008; Cooper, Blaser, & Brown, 1967; Peoples & Koch, 1979). Soils in the southern Coastal Plains are often low in plant?available K due to low cation exchange capacity (CEC; Sonon, Kissel, & Saha, 2014). Consequently, at least an additional 280 kg K2O ha−1 is recommended annually to maintain established alfalfa in Coastal Plain soils that contain <78 kg K2O ha−1 based on routine soil testing (Kissel & Sonon, 2011). Since alfalfa is a luxury consumer of K, splitting K applications across the season is recommended to better distribute nutrient uptake through the growing season to improve K utilization. With the increasing cost of fertilizer, producers are looking to minimize inputs. Potash averaged US$0.42 kg−1 at the time of this publication and the price trend was increasing (Knorr, 2019). Even though applying K below the recommended level will reduce input costs for producers, it is important to evaluate the consequences this could have on alfalfa forage production and persistence.
Harvest timing also influences alfalfa stand health and productivity. Too frequent defoliation can damage the stand and decrease longevity (MacLeod, Kunelius, & Calder, 1972) which diminishes yield potential. Alternatively, longer harvest intervals decrease digestibility and nutritive value (Palmonari, Fustini, Canestrari, Grilli, & Formigoni, 2014). Cold climates and dormancy ratings of alfalfa common in the northern United States limit seasonal production; however, in the southern United States, extreme high temperatures with little rainfall tend to slow alfalfa growth in the summer months (Brown, Hoveland, & Karnok, 1990). Later frost dates and semi? to non?dormant alfalfa varieties can enable harvesting of alfalfa from early spring (March) into fall (November) in the region. Forage production can extend beyond this range with mild winters. If biomass is not removed, carbohydrate reserves can be reduced due to shading of regrowth, thereby hindering photosynthetic ability (Teixeira, Moot, Brown, & Fletcher, 2007).
The variable climate of the Southeast region and sandy soils throughout the Coastal Plains present challenges for producers, especially those who do not apply sufficient K fertilizer as recommended by soil test analysis. While there are alfalfa stands that have survived over a decade in the region, the average alfalfa stand in our environment, under proper management, is 4 yr. In the Deep South Coastal Plains region, it has been proven that the alfalfa stand is profitable if it survives 3 yr (Prevatt, 2019). The objective of this trial was to determine the impact of reduced rates of K fertilization on cultivar Bulldog 805 alfalfa managed under different harvest regimes on forage yield, stand persistence, and nutritive value when grown in the southern Coastal Plains.
Core Ideas
Potassium fertilization and harvest timing influence stand life of alfalfa.
Objective: determine optimum K fertility and harvest timing for alfalfa in the South.
Reduced rates of K did not positively influence aboveground plant response.
Harvesting alfalfa at later maturity increases yield but decreases nutritive value.
Current alfalfa recommendations should be maintained to optimize yield and quality.
2 MATERIALS AND METHODS
2.1 Description of research site
This study was conducted from May 2017 until November 2018 at the University of Georgia's Tifton Campus in Tifton, GA (31°30′ N, 83°31 W; 110 m elevation). The research site was nearly level (<2% slope) and was composed of Tifton loamy sand soils (fine?loamy, kaolinitic, thermic Plinthic Kandiudults). Average temperatures for the 2?yr study were similar and consistent with the 100?yr average for the location (Figure 1a; UGA Automated Environmental Monitoring Network, 2018). Notably, temperatures never dropped to a level low enough to cause significant stand damage due to winter kill (–3°C). The 2018 growing season received above?average precipitation throughout the season, whereas the 2017 season was near average (Figure 1b).
image
FIGURE 1
Open in figure viewer
(a) Monthly mean air temperature and (b) total precipitation for 2017, 2018, and the 100?yr average in Tifton, GA. Data from UGA Automated Environmental Monitoring Network (UGA?AEMN, 2018)
2.2 Experimental design and forage management
Cultivar Bulldog 805 alfalfa (Athens Seed Co.) was planted into an established stand of cultivar Tifton?85 bermudagrass in November 2015 with a 2007 Pasture Pleaser Tye drill (AgCo) at 22 kg ha−1 pure live seed at a soil depth of 1.27 cm and 35.6 cm row spacing. The bermudagrass was terminated prior to data collection in April 2017 using clethodim at a rate 168 g a.i. ha−1 (Select MAX; Valent, Valent USA Corporation). Grass weeds were controlled throughout the study using clethodim applied six times at a rate of 168 g a.i. ha−1 and Sethoxydim (Poast; BASF Corporation) applied once at a rate of 0.34 kg a.i. ha−1.
Plots were arranged in a split plot design with four replications. Harvest timing treatments (whole plot) included bud, 10, 30, and 50% bloom stage (Sheaffer, Lacefield, & Marble, 1988). All whole plot treatments were randomly assigned within each block. Potassium fertilization treatments (subplot) included seasonal rates of 0, 67, 101, 134, and 168 kg ha−1 K2O (muriate of potash; 0–0–60). Each subplot factor was randomly applied within each whole plot factor. This resulted in 80 plots that measured 2.8 m2 each. Seasonal potash rates were split and hand?applied three times throughout the season: before the first cutting (Spring), 10 d after the second cutting (Summer), and 10 d after the second to last cutting (Fall) to allow for plant regrowth after harvesting and support nutrient uptake. All K fertilization rates were lower than current University of Georgia recommendations for established alfalfa maintenance and production in Coastal Plain soils with low (<78 kg K2O ha−1) soil test K (Kissel & Sonon, 2011).
All plots were fertilized with P and micronutrients (Mo and B) in accordance with soil test recommendations. These nutrients were applied at the initiation of each growing season and totaled 84 kg ha−1 P2O5 (monoammonium phosphate 12–61–0, Haifa; Haifa North America), 5.8 kg ha−1 B (B10%; CNI Liquid, CNI AgriMinerals), and 0.23 kg Mo ha−1 Mo (Mo10%; Chem?Plex Liquid, CNI AgriMinerals). Throughout the study pH was near neutral at both the topsoil and subsoil levels (pH 6.42 and 5.98, respectively).
Scouting for insect pests occurred weekly throughout the growing season and insecticide applications occurred as necessary following management recommendations for alfalfa in the region.
2.3 Response variables
Plots were visually assessed prior to each harvest to determine percent alfalfa ground cover and percent bloom. Growth stages were confirmed based on the procedure described by Mueller and Fick (1989). Fifty shoots were collected from each plot by hand?harvesting random shoots at 2.5 cm from the ground. To determine leaf/stem ratio, all leaves were removed from 25 stems before leaves and stems were placed in separate paper bags (stems included apical buds and flowers) to be dried. The remaining 25 stems were placed in a separate paper bag for drying.
Plots were harvested with a 1?m Swift Current Forage Plot Harvester IV (Rem Manufacturing; Thompson, 1972). Forage harvested from each plot was placed onto a 1.4?m2 tarp and hung from a tripod with a hanging scale (Intercomp) to record plot yield. A subsample was taken from each plot for dry matter (DM) determination. All samples were dried at 60°C for 72 h before they were weighed. Samples were ground in a Wiley Mill (Thomas Scientific) to pass through a 1?mm screen and prepared for lab analysis.
Mass shoot−1 and shoots m−2 were calculated using Equations 1 and 2, respectively.
Massshoot−1(g/shoot)=Massof25shoots(g)25
Massshoot−1g/shoot=Massof25shootsg25
(1)
Shootsm−2(shoots)=Yield(g/m2)Massshoot−1(g/shoot)
Shootsm−2shoots=Yieldg/m2Massshoot−1g/shoot
(2)
Stand density was assessed before the first and at the last harvest each year by counting individual crowns and associated stems within three, 0.1?m2 quadrats per plot.
Soil samples were taken within each plot to determine nutrient status at the topsoil (0–15 cm) and subsoil (15–30 cm) level at the initiation, midpoint, and conclusion of the evaluation. Soil samples were shipped to the University of Georgia Agricultural and Environmental Services Lab in Athens, GA, and analyzed using the Mehlich?1 procedure (Mehlich, 1953).
Tissue samples were collected from each plot at the conclusion of each season by harvesting the top 15 cm from the top of all plants within a plot. These samples were also dried at 60°C for 72 h, weighed, and ground through a Wiley Mill (Thomas Scientific) to pass a 1?mm screen. Samples were shipped to the University of Georgia Agricultural and Environmental Services Lab (Athens, GA) and analyzed for plant tissue K content.
After the final harvest in Year 2, a destructive harvest was performed on all plots to collect roots for carbohydrate analysis. A 0.5 by 1?m steel rectangular frame was manufactured to attach to a three?point hitch on a tractor. The bottom edge of the frame was sharpened to create a blade to cut through the soil. Approximately 227 kg was added as tractor weights to the back of the frame to maintain depth within the soil. Starter trenches were dug 1?m deep at the beginning of each row. The cutter bar was placed in the trench and dragged to the end of the row to cut the roots. Roots were excavated from three, 0.1?m2 quadrats after associated crowns and stems were counted. Roots were immediately washed in water and transported on ice to be hammer?milled (No. 10 Hammer Mill, CS Bell Company), vacuum sealed in a FoodSaver bag (Sunbeam Products), and frozen at –18°C. Samples were later freeze?dried and ground to pass a 1?mm screen in a Wiley Mill and then subsequently ground through a 1?mm screen in a Cyclone Sample Mill (Model 3010?030; UD Corporation). Ground root samples were shipped to Cumberland Valley Analytical Lab (Waynesboro, PA) for nonstructural carbohydrates (NSC), ethanol?soluble carbohydrate (ESC) and starch concentrations using the procedures outlined in Hall (2009) and Dubois, Gilles, Hamilton, Rebers, and Smith (1956).
2.4 Forage analyses
All forages samples were analyzed using near?infrared spectroscopy (NIRS; NIRSystems 6500 FOSS) with the 2018 legume hay equation developed by the NIRS Forage and Feed Testing Consortium (Hillsboro, WI) to determine crude protein (CP), in?vitro true dry matter digestibility (IVTDMD), and parameters associated with the calculation of total digestible nutrients (TDN) as seen in Equation 3 (NRC, 2016)
TDN=(1.0787×CP)+(1.327×Fat)+(0.4208×NDF)+(0.9689×NFC)
TDN=1.0787×CP+1.327×Fat+0.4208×NDF+0.9689×NFC
(3)
where CP = crude protein; Fat = fat; NDF = neutral detergent fiber; NFC = nonfibrous carbohydrates (NDF × 0.93) (NRC, 2016).
A subset of samples (n = 152) was used to validate the NIRS results via comparison of samples analyzed using wet chemistry. Samples were analyzed for DM and ash content (AOAC, 2000); NDF (Van Soest, Robertson, & Lewis, 1991) and ADF (AOAC, 2000) using an ANKOM 2000 analyzer (ANKOM Technology; Mertens, 2002); in vitro true dry matter digestibility as described by Promerleau?Lacasse, Seguin, Bélanger, Lajeunesse, and Charbonneau (2019); and CP content using a LECO combustion analyzer (model FP628, Leco Corporation).
2.5 Statistical analysis
For statistical analysis, harvests were categorized into defined harvest periods of one complete cycle of all four growth stage harvests rather than a determined number of days between harvests (Table 1).
TABLE 1. Harvest date range for one cycle within year by period
Harvest period 2017 2018
1 5 May–15 May 6 Apr.–27 Apr.
2 1 June–15 June 17 May–31 May
3 29 June–24 July 11 June–26 June
4 7 Aug.–29 Aug. 9 July–24 July
5 5 Sept.–5 Oct. 30 July–24 Aug.
6 6 Oct.–5 Nov. 28 Aug.–25 Sept.
7 – 1 Oct.–9 Nov.
Data were analyzed by restricted maximum likelihood using PROC MIXED in SAS v9.4 (SAS Institute). The Kenward–Rodgers adjustment was used to correct the denominator degrees of freedom and ensure appropriate standard errors and F statistics for each tested model. The Autoregressive (1) covariance structure was determined to be the best fit for these models based on the lowest Bayesian's Information Criterion (Littell, Milliken, Stroup, Wilfinger, & Schabenberger, 2006). Fixed effects for each response model included K treatment, harvest treatment, and K treatment × harvest treatment and were tested within harvest period. Year, block, block × K treatment, and block × harvest treatment were considered random effects for these models. Fixed effects for each belowground responses model were K treatment and harvest treatment. Data were analyzed within soil depth and year for soil K level. Random effects included year (where available), block, block × K treatment, and block × harvest treatment. All means were compared using the LSMEANS procedure with Tukey–Kramer adjustment. Differences were considered significant at P ≤ .05.
3 RESULTS AND DISCUSSION
3.1 Aboveground plant responses to potassium fertilization
Evaluated levels of K fertilization had no significant effect on alfalfa yield (P > .59), mass shoot−1 (P > .21), leaf/stem (P > .24), shoots m−2 (P > .47), percent alfalfa ground cover (P > .09), or stand density via crown and stem counts (P > .10) (data not shown). Additionally, K treatment had no effect on CP (P = .64), IVTDMD (P = .59), or TDN (P = .49) (data not shown). Finally, K fertilization rate had no significant effect on K content of tissue samples (P = .99) (data not shown).
The reduced rates of K used in this study were all below current University of Georgia recommended levels of potash applications for alfalfa management. Soil K levels were in the medium range across all plots (avg. = 91.8 kg ha−1) at project initiation for successful growth of alfalfa, therefore we did not see a significant response in plant factors.
3.2 Aboveground plant responses to harvest timing
Differences for mass shoot−1 occurred in harvest periods 2, 3, 5, 6, and 7 (P < .01), but no clear trends existed (Table 2). Harvest timing only affected shoots m−2 in harvest periods 3, 5, and 7 (P < .05), but again no clear trend existed (Table 2). As a result, alfalfa cover did not differ any of harvest periods 2 through 6 (P > .10; Table 2). There was a difference in alfalfa cover in harvest periods 1 and 7 such that treatments harvested at later stages generally had greater cover (P < .04; Table 2).
TABLE 2. Effect of harvest timing (based on growth stage) within harvest period on alfalfa plant responses in Tifton, GA; data pooled over year and block
Harvest period
Growth stagea 1 2 3 4 5 6 7
mass shoot−1 (g shoot−1)
Bud 1.4 0.9bb 1.1ab 1.1 0.8b 1.0b 1.1a
10% Bloom 1.2 0.1ab 1.1ab 1.1 0.9a 0.9b 0.9b
30% Bloom 1.3 1.1a 1.2a 1.0 1.1a 1.2a 1.1a
50% Bloom 1.5 1.2a 1.0b 1.0 0.9ab 1.3a 0.1ab
SE 0.10 0.04 0.05 0.04 0.05 0.06 0.05
P value .34 <.01 <.01 .38 <.01 <.01 <.01
shoots m−2
Bud 110 117 139ab 116 125ab 139 120ab
10% Bloom 142 151 171a 112 98b 146 138ab
30% Bloom 129 101 112b 147 106ab 156 95b
50% Bloom 126 105 150ab 167 152a –c 158a
SE 15.0 25.2 21.3 24.7 14.8 67.7 17.5
P value .21 .32 .05 .07 .04 .41 .05
alfalfa cover, %
Bud 61b 76 81 80 66 78 77ab
10% Bloom 75ab 78 81 78 77 75 68b
30% Bloom 77a 73 75 78 69 74 80ab
50% Bloom 78a 76 78 76 78 80 86a
SE 8.0 5.3 2.7 7.7 3.9 5.6 3.7
P value .03 .83 .10 .91 .13 .63 .04
aAs defined by Mueller and Fick (1989).
bMeans within a column followed by the same letter are not significantly different at (P ≤ .05)
cData points removed due to significant lodging (–).
Typically, with increased shoots m−2 and ground cover, shoot mass decreases per plant to accommodate the same space with more shoots (Teixeira et al., 2007). Berg et al. (2005) found an increased mass shoot−1 associated with increased alfalfa yield, which was not seen in this study. Perhaps differences in mass shoot−1 were not seen because alfalfa was planted on 36?cm?wide rows and other authors found that wide row spacing may allow for greater light interception at crown level and could facilitate greater shoot production (Ventroni, Volenec, & Cangiano, 2010). Timing of potash applications may have caused the sporadic treatment differences observed in Table 2. Alfalfa harvested at 10 or 50% bloom produced greater cumulative yield than alfalfa harvested at the bud stage; alfalfa harvested at 30% bloom was intermediate to and not different from any of the other growth stages (P < .01; Table 3). Yield differences did exist in two harvest periods, but again clear trends were not observed (Table 3). In harvest period 3, yields from plots harvested at 10% bloom were greater than 30% (P = .03). In harvest period 5, alfalfa harvested at 50% bloom yielded more than that harvested at bud and 10% but was not different from the alfalfa harvested at 30% (P < .01). Greater lodging in the 30 and 50% bloom stage could have lowered the ability to capture total yield with the harvester. Brown et al. (1990) reported increased leaf senescence in later maturity stages up to 50% as a result of the hot conditions in the Coastal Plains, which indicated no yield advantage for 50% over 10% bloom.
TABLE 3. Effect of harvest timing (based on growth stage) within and across harvest period on alfalfa yield in Tifton, GA; data pooled over year and block
Harvest period
Growth stagea 1 2 3 4 5 6 7 Cumulative (per season)
kg ha−1
Bud 1,242 848 1,328abb 1,068 744b 1,147 1,199 14,840b
10% Bloom 1,423 1,325 1,612a 1,043 774b 1,103 958 16,761a
30% Bloom 1,490 937 1,125b 1,279 908ab 1,006 953 15,796ab
50% Bloom 1,497 1,055 1,245ab 1,410 1,142a 948 1,359 16,803a
SE 80.6 174.4 158.4 182.2 70.6 425.3 121.6 465
P value .07 .23 .03 .08 <.01 .63 .05 <.01
aAs defined by Mueller and Fick (1989).
bMeans within a column followed by the same letter are not significantly different at (P ≤ .05).
As expected, growth stage affected all forage nutritive value parameters. Although some variability was seen throughout the growing season, CP, TDN, and IVTDMD generally decreased as maturity increased (Table 4), congruent with the results reported by several other authors (i.e., Gasser et al., 1969; Kalu & Fick, 1983). Consequently, forage fiber components, NDF, ADF, and lignin increased (data not shown), similarly to Kallenbach, Nelson, and Coutts (2002) and Sheaffer et al. (2000) . Finally, end of trial data from tissue samples determined that growth stage affected K content (P < .01) in that bud stage was deficient (19.7 g kg−1), however all other treatments were within the sufficiency range of 22.5–34.0 g kg−1 for alfalfa (32.8, 29.3, and 29.6 g kg−1) as reported by Undersander et al. (2011) and Plank and Kissel (2019).
TABLE 4. Effect of harvest timing (based on growth stage) within and across harvest period on alfalfa nutritive value parameters in Tifton, GA; data pooled over year and block
Harvest period
Growth stagea 1 2 3 4 5 6 7
crude protein, g kg−1
Bud 247ab 221c 229a 227a 240 244a 256
10% Bloom 223b 229bc 234a 224a 239 237ab 270
30% Bloom 221b 257a 212b 226a 235 231b 270
50% Bloom 206c 238b 210b 201b 232 242ab 272
SE 2.4 2.1 3.7 3.8 2.6 2.3 4.9
P value <.01 <.01 <.01 <.01 .18 <.01 .13
total digestible nutrients, g kg−1c
Bud 685a 680a 651a 666a 654b 703a 724
10% Bloom 665b 680a 646a 646ab 674a 677b 722
30% Bloom 670ab 686a 601b 642bc 661b 663b 722
50% Bloom 660b 657b 640a 622c 676a 677b 708
SE 5.1 4.1 5.0 6.1 3.3 4.4 4.7
P value <.01 <.01 <.01 <.01 <.01 <.01 .13
in vitro true dry matter digestibility, g kg−1
Bud 827a 811b 787a 812a 816ab 846a 878a
10% Bloom 797b 814ab 790a 797a 823a 821b 854b
30% Bloom 801b 823a 759b 801a 812b 800c 845b
50% Bloom 788b 799c 779a 768b 822ab 833b 853b
SE 4.7 3.9 6.3 6.9 3.7 4.4 5.3
P value <.01 .02 <.01 <.01 .11 <.01 <.01
aAs defined by Mueller and Fick (1989).
bMeans within a column followed by the same letter are not significantly different at (P ≤ .05).
cTDN: TDN = (1.0787 × CP) + (1.327 × Fat) + (0.4208 × NDF) + (0.9689 × NFC) (NRC, 2016).
3.3 Belowground responses to potassium fertilization
At the beginning of the 2?yr study, K levels in all plots were in the medium range at both the topsoil and subsoil levels (avg. = 91.8 kg ha−1). Across the 2?yr study, K content differed by treatment in both topsoil and subsoil levels (P < .01; Table 5). After initiation, topsoil K levels were lowest in the untreated control (P < .01) throughout the study as all other treatments increased with K fertilization At the study conclusion, subsoil K levels were lower in the untreated control than most K fertilization treatments except the 67 kg ha−1 treatment, which was different from the 168 kg ha−1 treatment(P < .01). Subsoil K levels decreased in all treatments throughout the evaluation, however at the study conclusion we begin to see a separation in soil K content between our lower and higher rates, with only the highest rate (168 kg ha−1) being able to better maintain a subsoil K level in the medium range (80–190 kg ha−1). Had this project continued for an additional year, the authors feel that this separation would continue to become more prominent. Additionally, it should be noted that sandy soils have a high potential for nutrient leaching of K, and K has more influence on stand longevity than establishment (Undersander et al., 2011; Lacefield et al., 2009), therefore the decrease in subsoil K content as reported in this study would inhibit nutrient availability to the plant in the root zone, thus impacting long?term stand success. Potassium fertilization did not elicit responses in any root carbohydrate analyses (including starch, NSC, ESC; data not shown; P > .25).
TABLE 5. Effect of K fertilization at experiment initiation, midpoint, and conclusion on topsoil (0–15 cm) and subsoil (15–30 cm) K content in alfalfa grown in Tifton, GA; Tifton loamy sand soils (fine?loamy, kaolinitic, thermic Plinthic Kandiudults)
Timepoint
Soil level K treatment Initiation Midpoint Conclusion
kg ha−1 kg K ha−1
Topsoil 0 95.2 82.2aa 59.4a
67 116.9b 128.4b
101 127.5b 133.1b
134 127.7b 138.0b
168 140.4b 155.7b
SE 11.8 12.4
P value .02 <.01
Subsoil 0 88.3 52.2 36.3a
67 68.5 56.4ab
101 74.1 61.2bc
134 71.0 70.1bc
168 77.3 81.5c
SE 8.4 6.5
P value .13 <.01
aMeans within a column followed by the same letter are not significantly different at (P ≤ .05).
3.4 Belowground responses to harvest timing
Root carbohydrate analyses determined that harvest timing affected starch and NSC content of roots (P < .01; Table 6). Both starch and NSC declined with increasing maturity stage at harvest with bud stage being the highest and 50% bloom treatment being the lowest for both analyses (P < .01). This was likely a consequence of carbohydrate remobilization within the plant. As alfalfa maturity reaches the 50% bloom stage, shoot growth from new crown buds begins, this is supported by carbohydrates that are remobilized from taproot storage, thus causing a dip in root carbohydrate reserves (Brown, Pearce, Wolf, & Blaser, 1972; Heichel, Delaney, & Cralle, 1988). Harvest timing did not influence soil K (data not shown; P > .41).
TABLE 6. Effect of harvest timing (based on growth stage) on root starch and non?soluble carbohydrates of a three?year old alfalfa stand grown in Tifton, GA; data pooled over block
Growth stagea Starchb NSCb
g kg−1
Bud 193ac 316a
10% Bloom 164ab 303a
30% Bloom 150bc 292a
50% Bloom 113c 242b
SE 14.0 16.2
P value <.01 <.01
aAs defined by Mueller and Fick (1989).
bAs described in Hall (2009) and Dubois et al. (1956).
cMeans within a column followed by the same letter are not significantly different at (P ≤ .05).
4 CONCLUSION
Reduced rates of K fertilization did not influence any aboveground plant responses beyond the untreated control in this experiment. Harvesting at later maturity stages tended to increase yield, but also increased fiber content and lowered nutritive value. To optimize both alfalfa yield and persistence, current recommendations for alfalfa harvest timing and K fertilization should be maintained in the southern Coastal Plains. These include harvesting on a 28? to 35?d interval once alfalfa plants have reached the 10% bloom stage, fertilizing with at least the minimum amount of potash recommended by annual soil tests, and dividing the fertilizer into at least three applications during the season (Hancock et al., 2009). Future studies should focus on the long?term economic and agronomic impact of “cutting corners” by fertilizing below the recommended rate of potash in established alfalfa stands.
ACKNOWLEDGMENTS
The authors would like to thank the USDA?NIFA?Alfalfa Forage Research Program grant no. 2016?70005?25653 and the 2018 U.S. Alfalfa Farmer Research Initiative Checkoff for funding of this project and support from the University of Georgia. The authors also thank Zach Collins, Shauni Nichols, Melissa Tawzer, and the beef?forage graduate students and student workers at the University of Georgia Tifton Campus.