A proper P nutrition is essential for alfalfa plant survival and its supply is often necessary to reach maximum stand development, productivity, and persistence (Jung and Smith, 1959; Berg et al., 2005, 2007). Phosphorus is also involved in the nitrogen fixation process since it contributes to increase both the size and number of Rhizobia nodules (Azcόn et al., 1988). Yield responses to P are typically most evident at the first harvest of the season; however, the efficiency of P fertilization may be influenced by application timing (James et al., 1995). In northern Italy, P is commonly applied annually in early winter (December) and the current fertilization rate recommendations range from 40 to 50 kg ha–1 yr–1 of P2O5 (Parrini and Bonari, 2002). Moreover, alfalfa yield, forage nutritive value, and stand density are strongly affected by K nutrition (Wolf et al., 1976; Kafkafi et al., 1977; Lanyon and Smith, 1985). Nevertheless, alfalfa can absorb more K than needed for maximum growth through a behavior called “luxury” consumption, which generally leads to a reduction in protein, Ca, Mg, and Na (Lanyon and Griffith, 1988; Pant et al., 2004). High concentrations of K (e.g.,>3% dry wt.) in alfalfa may be undesirable for livestock nutrition because the reduced concentration of Ca and other nutrients impact animal utilization of alfalfa forage (Meyer and Matthews, 1995). In southern Europe, K fertilizer is typically applied once a year as a single application during the winter, at a rate of 200 to 300 kg ha–1 yr–1 of K2O (Lloveras et al., 2001; Parrini and Bonari, 2002). However, split applications of K can reduce alfalfa plants’ tendency for luxury consumption and are considered more efficient than a one-time application of a large amount of K (Kresge and Younts, 1962; Kafkafi et al., 1977). In particular, Kafkafi et al. (1977) recommended to use split applications of K in soils having low cation exchange capacity (e.g., 5.5 cmolc kg–1).
Research investigating the effects of P and K on the agronomic performance of alfalfa suggests that the interactions between nutrients should not be underestimated. Lissbrant et al. (2010) demonstrated that, under various soil conditions, lower yields were obtained by unbalanced P and K fertilization compared to zero fertilization, confirming the interdependent nature of these nutrients. Still, Lanyon and Griffith (1988) reported that fertilizer efficiency varied across locations because the plant response to fertilization depends on many environmental factors, such as soil nutrient availability, soil status, and climate conditions. According to this, alfalfa fertilization programs should be customized on the basis of local conditions, which may also change over time.
For decades in most regions with intensive agricultural systems, producers increased fertilization inputs and especially P rates to achieve higher yields, often leading to soil nutrient accumulation (Haden et al., 2007; Delgado and Scalenghe, 2008; Pizzeghello et al., 2011). However, there is little information in the current body of scientific literature on fertilization effects from areas where soil P and/or K are not limiting. Barbarick (1985) reported that high K application to a high K soil increased irrigated alfalfa yield only after 4 yr, while no such effect was measured by Lloveras et al. (2001). To reduce production costs and to prevent further accumulation of nutrients in the soil, the alfalfa fertilization programs should be tailored on the basis of the changes in soil nutritional status. Therefore, the objective of this study was to investigate the response of alfalfa yield and nutritive value to various combinations of P and K rates in a soil lacking nutrient deficiency.
MATERIALS AND METHODS
A field trial was performed during three growing seasons (2005–2007) at the experimental farm of Padova University in Legnaro, Italy (45°20′ N, 11°58′ E; elevation 8 m). The zone is characterized by a humid subtropical climate with an average annual temperature of 12.3°C and 820 mm total rainfall. Monthly mean air temperatures and monthly precipitation during the investigation period and the 45-yr monthly averages are reported in Fig. 1. The first two growing seasons were characterized by low precipitation in June and above-average precipitation during late summer. Monthly air temperatures were generally similar to the long-term averages, with the exception of winter 2006–2007, during which temperatures were extraordinary high (Fig. 1). Soil at the site was an Oxyaquic Eutrudept, coarse-silty mixed, mesic (Morari, 2006), having a silt loam texture, high content of available P, and medium content of exchangeable K (Table 1). The total cation exchange capacity was 21.6 and 17.2 cmolc kg–1 for the 0- to 46- and 46- to 73-cm depths, respectively (Giupponi and Dibona, 2004). On 30 Mar. 2005, alfalfa cultivar Delta was seeded at a rate of 29 kg seed ha–1 in rows spaced 20 cm apart. The cultivar used in this study was nondormant (fall dormancy rating = 8.5) and it was selected on the basis of its common use in the area of investigation (Rimi et al., 2010).
|Parameter||Method of determination||Value|
|Sand, % (0,05–2 mm)||gravimetric||18.0|
|Silt, % (0002–0,05 mm)||gravimetric||65.9|
|Clay, % (<0002 mm)||gravimetric||16.0|
|Organic matter, %||Walkley and Black||2.28|
|Total N, g kg–1||Kjeldahl||0.125|
|Total P, ‰ P2O5||mineralization via HNO3 and H2SO4||1.9|
|Available P, mg kg–1||Olsen||38.2|
|Exchangeable K, mg kg–1||buffered BaCl2||178.0|
|C/N ratio||Organic C ÷ Organic N||10.7|
A total of nine treatments deriving from the factorial combination of three fertilization rates of P and three of K were compared in a randomized complete block design with three replicates. The fertilization treatments included: 0, 100, or 200 kg ha–1 of P2O5, and 0, 300, or 600 kg ha–1 of K2O. The first fertilization was applied according to the treatments before seeding in March 2005, and subsequent applications were made in December of 2005 and 2006. The fertilizers were distributed as a single application each year, according to the common alfalfa management in northern Italy. Irrigation was limited to the establishment phase and plot size was 8 m2 (1.6 by 5 m), of which 6 m2 were used for sampling. Plots were harvested using a sickle-bar harvester at a 5-cm height four times from June through October 2005, seven times from May through October 2006, and seven times from April through September 2007. The first harvest of each growing season was made at the appearance of new crown shoots, while the subsequent harvests were made at an early bud phenological stage (Kalu and Fick, 1981).
At each harvest, canopy height was measured at four randomly selected locations within each plot and the values were averaged to obtain one value. In addition, all stems from 30 cm of two rows of each plot were counted to determine stem density (stem m–2). After harvest, the fresh herbage was weighed in the field and 500-g subsamples were collected and dried for 36 h at 65°C to determine dry matter (DM) yield. Subsequently, samples were analyzed for crude protein (CP), ether extract (EE), neutral detergent fiber (NDF), acid detergent fiber (ADF), acid detergent lignin (ADL), and ash concentration via calibrated near infrared reflectance spectroscopy (NIRS) (model 5000; NIRSystems, Silver Springs, MD). Calibration samples were subjected to conventional chemical analysis for CP (AOAC, 1995), NDF, ADF, ADL (Goering and Van Soest, 1970), EE (AOAC, 1990), and ash (Pearson, 1976). Spectra data analyses were performed using WinISI II software package (version 1.5; Infrasoſt International, Port Matilda, PA) and calibration statistics are presented in Table 2.
|Neutral detergent fiber||1.90||0.88|
|Acid detergent fiber||1.96||0.86|
|Acid detergent lignin||1.01||0.66|
In October 2007, three soil cores measuring 4.5 (diam.) by 20 cm (depth) were randomly collected on each plot, for determination of available P and exchangeable K at the end of the research period. The cores were divided into three sections representing the 0- to 5-, 5- to 10-, and 10- to 20-cm soil layers; and weighted averages were calculated for the 0- to 20-cm depth. The forage concentrations of P and K were determined using prediction equations developed by the NIRS Forage and Feed Testing Consortium (Hillsboro, WI). Calibration statistics were the following: P, standard error of cross-validation in modified partial least square regression (SECV) = 0.04 and r2 = 0.67; K, SECV = 0.27 and r2 = 0.88 (P. Berzaghi, personal communication, 2012). Subsequently, soil balances of P and K were calculated by subtracting the total forage uptake from the initial soil contents and the cumulative supply due to fertilization.
The data were subjected to analysis of variance using SAS Proc Mixed (version 9.2; SAS Institute, Cary, NC) to determine the main effects of P, K, and their interaction. Fertilization treatments showed little differences at individual harvests (data not shown), consequently the yield data were summarized as annual totals, whereas the nutritive values were weighted by yield values (Kallenbach et al., 2002; Rimi et al., 2012). Replicates were considered random and P, K, and their interaction were considered fixed. When appropriate, years or soil sampling depths were included in the statistical model as repeated measurements. Fisher’s Protected LSD test was used at the 0.05 probability level to identify significant differences between means.
RESULTS AND DISCUSSION
For DM yield, P and all interaction terms including P were not significant (Table 3), which was probably due to the high level of available P in the soil at seeding (Table 1). Therefore, the application of P did not interact with K and appeared basically needless to improve yield; however, an increase of fertilization from 100 to 200 kg ha–1 of P2O5 was not detrimental to yield. The analysis of variance for DM yield showed significant K and year main effects, and their interaction (Table 3). Observing the interaction between K and years, there were differences among K rates only in the second and third year of investigation (Table 4). In 2006, the unfertilized plots had lower DM yield compared to those that received K2O, which showed no differences between each other. In 2007, the difference in DM yield between unfertilized plots and fertilized plots further increased, and a significant difference was observed between the 300 and 600 kg ha–1 K2O treatments (Table 4). These findings corroborated the results of other studies that reported a positive influence of K on alfalfa yield even in soils with high exchangeable K levels (Markus and Battle, 1965; Kafkafi et al., 1977; Barbarick, 1985). In the average of 3 yr, the increase of K2O from 300 to 600 kg ha–1 did not provide any increment of DM yield (Table 4), suggesting that the application of 300 kg ha–1 was sufficient to maximize yield. These recommendations appeared in disagreement with previous studies from Indiana, where maximum levels of K fertilization increased alfalfa yields on each year during a 5-yr investigative period (Berg et al., 2005, 2007). The different responses observed between the studies could be attributed to contrasting initial soil nutritional status for the two sites and/or the levels of P and K fertilization imposed.
|Canopy height||Stem density||Forage nutritive values
|P × K||ns||ns||ns||ns||ns||ns||ns||ns||ns||ns||*|
|P × yr||ns||ns||ns||ns||***||ns||ns||ns||ns||ns||*|
|K × yr||***||***||ns||***||ns||ns||ns||ns||***||ns||**|
|P × K × yr||ns||ns||ns||**||ns||ns||ns||ns||ns||ns||ns|
|kg ha–1 yr–1|
|Dry matter yield, Mg ha–1|
|Canopy height, cm|
|Ash, g kg–1 dry wt.|
|K, g kg–1 dry wt
Similarly to DM yield, the canopy height was not affected by P, while it was influenced by K, years, and their interaction (Table 3). The canopy height responded to K treatments and was related to the environmental differences that occurred between years consistently to DM yield (Table 4). The stem density was not influenced by P or K (Table 3), but a decrease in stem density was observed over the three experimental years, with stem m–2 diminishing from ≈1100 to ≈550 (data not shown). This could be attributed to stand thinning and a natural physiological reduction in the number of stems per plant over time (Hall et al., 2004). Therefore, additional K application possibly affected DM yield through stem elongation rather than stem density.
While the effect of years was significant for all forage nutritive values, changes in consequence of fertilization treatments were limited to CP, EE, ash, P, and K (Table 3). The analysis of variance for CP revealed a significant three-way interaction among P, K, and years (Table 3). In the establishment year (2005), all plots had similar CP, with the exception of those receiving 200 kg ha–1 P2O5 and 0 kg ha–1 K2O, which had lowest content of CP (Table 5). In the subsequent year, the plants unfertilized with K were in the top statistical group, while those receiving 600 kg ha–1 K2O had lowest CP, regardless of P fertilization (Table 5). In 2007, plots fertilized with 600 kg ha–1 K2O had again lowest content of CP when combined with the 100 and 200 kg ha–1 P2O5 rates (Table 5). In general, the negative effect of high K applications on CP could be ascribed to a decrease of leaf/stem ratio, possibly due to increased stem length as suggested by canopy height (Table 4). Also, lower CP could be explained by a diminution of N content per leaf area as the result of higher competition for light in consequence of the greater plant height and yield (Sheafler et al., 1986; Lemaire et al., 1991). Moreover, there was a significant interaction between P fertilization rates and year on EE (Table 3), however the differences among treatments were of little biological importance (data not shown). There were significant differences in ash content among K fertilization rates and a significant K × year interaction (Table 3), due to a continued increment of ash over years with the increase of K rate (Table 4).
|kg ha–1 yr–1||g kg–1 dry wt.|
The accumulation of K in the tissues was promoted by K fertilization rates and was also influenced by the two-way interactions P × K, P × year, and K × year (Table 3). The forage concentrations of K increased as K2O fertilization passed from 0 to 300 and from 300 to 600 kg ha–1, although there was a difference in magnitude between years (Table 4). The increase of K concentration in the forage in response to high levels of K2O may be explained by a luxury consumption of this nutrient (Pant et al., 2004; Snyder and Leep, 2007). These results indicated that luxury consumption of K can occur even in soils having high cation exchange capacity. The forage content of K was positively correlated with ash (r = 0.761; P < 0.0001; n = 81), suggesting that high ash content in alfalfa grown on soils rich in K is due to luxury consumption. The forage content of K was influenced by P fertilization only in plants that had not received K2O, with forage K decreasing when P2O5 passed from 0 to 200 kg ha–1 (Table 6). The impact of P fertilization on the forage concentrations of K across years appeared unclear, with statistical differences among treatments noticeable only in 2006 (Table 6). Interestingly, P fertilization had no effect on the concentration of P in the forage, which was influenced only by the main effects of K and years (Table 3). In fact, the forage concentrations of P were lower for plots unfertilized (29 mg kg–1 dry wt.) compared to those fertilized with 300 and 600 kg ha–1 K2O (30 mg kg–1 dry wt.) (data not shown). These results reflected the lack of response of DM yield in consequence of P application and are likely due to the high availability of P in the soil at seeding.
|K2O, kg ha–1 yr–1
|kg ha–1 yr–1||mg kg–1 dry wt.|
At the end of the investigative period, the available P in the upper 20 cm of soil was still sustained, although differences occurred among P fertilization treatments (Tables 7 and 8). Moreover, P fertilization affected the P balance such that only the plots that received 200 kg ha–1 yr–1 offset the P removed by the canopy over the 3-yr period (Table 8). In addition, differences in P balances were noticed among the K fertilization treatments (Table 8), indicating that K nutrition may interact with P in determining P sorption and utilization. Soil exchangeable K measured at the end of 2007 on plots fertilized with 600 kg ha–1 K2O was significantly higher compared to plots that received 0 and 300 kg ha–1 K2O (Table 8). However, no biologically significant increases in soil K were noticed when the values measured after the final harvest (Table 8) were compared to the soil concentrations at seeding (178 mg kg–1; Table 1). These results indicated a high use of K by alfalfa even when it is supplied at high fertilization rates. The decrease of exchangeable soil K content observed in plots receiving 0 and 300 kg ha–1 K2O was substantiated by a negative soil balance for K (Table 8). Among the fertilization treatments, only the 600 kg ha–1 K2O rate was sufficient to offset the K removed by alfalfa over the 3 yr. The lack of yield differences between 300 and 600 kg ha–1 K2O until the third year (Table 4) suggested that 300 kg ha–1 yr–1 K2O can be applied for the first 2 yr in high K soils. As such, soil nutritional status should be monitored in the following year, to adjust the application rate and sustain alfalfa yields.
|Source||0–20 cm||0–20 cm||P||K|
|P × K||ns||ns||ns||ns||ns||ns|
|P × D||***||–||ns||–||–||–|
|K × D||ns||–||***||–||–||–|
|P × K × D||ns||–||ns||–||–||–|
|Fertilization treatment||Soil P||Soil K||P||K|
|P2O5, kg ha–1|
|K2O, kg ha–1|
At the end of the investigative period, soil concentrations of available P at different soil depths were influenced by P, soil sampling depths, and their interaction, while K had no effect (Table 7). A depletion of P at each soil layer was observed in unfertilized plots, while in plots receiving 100 and 200 kg ha–1 P2O5 depletion occurred at 5- to 10- and 10- to 20-cm layers (Fig. 2). However, the final P content at each soil layer, and for each P rate, was equal or higher than the amount of ≈22 mg kg–1, which is considered adequate to sustain 3-yr yield in a silt loam soil (Lanyon and Griffith, 1988). The high levels of available P observed in the upper layer for the 100 and 200 kg ha–1 P2O5 treatments (Fig. 2) were likely related to the scarce mobility of P in the soil (Havlin et al., 1999).
As with P, final soil concentrations of exchangeable K at different soil depths were influenced by K, soil sampling depths, and their interaction, while P had no effects (Table 7). Differences in soil K content among K rates occurred only in the upper soil layer, where the plots fertilized with 600 kg ha–1 K2O had higher K than other treatments (Fig. 3). The lack of differences in the deeper soil layers (5–10 and 10–20 cm) indicated that yield differences were most likely related to the K content of the upper soil layer. These results also suggested that the efficiency of alfalfa to remove K from the lower soil profiles was not affected by the amount of K available in the 0- to 5-cm depth. Still, it was interesting that soil exchangeable K of unfertilized control plots at each soil depths were lower than 150 kg ha–1, which is recognized as a limit to sustain alfalfa yield (Snyder and Leep, 2007). However, the results of this study contradicted the previous recommendations of Snyder and Leep (2007), since the initial soil K content of 178 mg kg–1 (≈ 230 kg ha–1) limited alfalfa performance in unfertilized plots.
In a soil lacking P and K deficiencies, an annual application of 300 kg ha–1 of K2O was more efficient in comparison with 600 kg ha–1 of K2O and no K fertilization. In fact, this fertilization schedule allowed to maximize yield and to preserve forage nutritive value over a period of three growing seasons. In contrast, the P fertilization had no impact on yield or forage nutritive value and it caused additional accumulation of this element in the soil over time, increasing potential environmental impacts. Although alfalfa plants evidenced a sustained K uptake during the growing season, high K application rates led to excessive accumulation of K in the upper soil layer. The results of this study suggest that the traditional fertilization practices for high-yielding alfalfa growing in soils having a high nutritional status should be judiciously reconsidered. To prevent excessive applications of P and K and possible luxury consumption of K, alfalfa producers are encouraged to perform soil tests on a regular basis and to schedule annual applications according to soil nutritional status.