Influence of sodium and potassium fertilization on the sodium concentration of timothy

Sodium (Na) concentration of forage crops grown in Finland, particularly that of timothy, is much lower than is recommended in the feed of cattle. A pot experiment was carried out on clay, loam and organogenic soils to find out the effect of Na application (0, 200 or 400 mg dm 3 of soil, one application) on the concentration of Na, K, Ca and Mg of timothy and the effect of K fertilization (0, 100 and 200 mg dnr3 for each three harvests) on the efficiency of Na application. Added Na elevated the Na concentration in all harvests on all soils. The magnitude of the effect (organogenic soils loam>clay) was opposite to the K supplying power of the soil. Potassium fertilization suppressed the effect of Na application substantially and Na concentration was elevated remarkably only when the K concentration of the plants fell to or below the deficiency level (approximately 15 g kg-1). According to a cation exchange experiment, nearly all added Na remained in the soil solution. Still, the apparent utilization ofadded Na remained below 4% on all soils, demonstrating the natrophobic nature of timothy. Sodium fertilization of timothy seems to be an ineffective way of increasing the Na content of forage at least on soils of a good K status or when applied with ample K fertilization.


Introduction
Sodium (Na) can substitute for potassium (K) in the biophysical functions of K in most plants, e.g. in regulating osmotic pressure in vacuoles, and to a limited extent in biochemical functions.e.g. in activation of enzymes.However, the abil- ity of Na to substitute for K varies greatly be- tween plant species (Flowers and Läuchli 1983).Most agronomically significant species like tim- othy, rye, corn and soybean are natrophobic.In these species Na cannot effectively substitute for K and the Na concentration of plants tends to be low.The average Na concentration of timothy in Finland (mean 0.047 g kg' 1 , Kähäri and Nissi-  nen 1978) is of the same order as that of the micronutrients Fe, Mn and Zn, while in other grass- es like cocksfoot (0.3 g kg 1 , Rinne et al. 1974), and particularly in ryegrass (0.8 g kg 1 , Jansson   1986), the Na concentration tends to be higher.
The Na concentration of cereals is relatively unimportant but that of pasture species affects the quality offodder used for animal production.Sodium is an essential mineral element for animals, and forage should contain Na 1.8-2g kg 1 of dry matter to supply milking cows with suffi- cient Na (NJF 1975, Smith and Middleton 1978,  Horn 1988).Thus, there is a big difference be- tween the requirement of cattle and the supply of Na from farm-produced fodder in Finland, and it has to be compensated with mineral supplements.Another imbalance of mineral elements in fodder is brought about by heavy K fertiliza- tion of leys leading to excessive K concentra- tion and lower than optimum concentration of other cations, especially magnesium (Mg), in herbage (Smith and Middleton 1978, Leigh et  al. 1988).The equivalent ratio K/(Ca+Mg) has been used as a criterion for forage grass quality; values below 2.2 are desired (Ettala and Kossila  1979).Substitution ofNa for K in plants has been shown to increase their Mg concentration (Nowa- kowski et al. 1974, Smith 1974, Smith et al.  1980, Mundy 1983).Sodium application can thus improve the feeding quality of herbage by increasing the concentration of Na and Mg and possibly by lowering the concentration of K. Elevated Na concentration of herbage may also increase the intake of fodder by cows (Horn  1988, Chiy et al. 1993), resulting in an increase in liveweight gain and in milk production (Chiy  et al. 1993).
The purpose of this study was to examine whether the Na concentration of timothy can be elevated by Na application and to find out the effects of Na application on the uptake of other cations on three different soils.The effect of K application on the efficiency of Na fertilization was also studied.The natrophobic nature of tim- othy and its poor response to added Na is evi- dent from the literature.Nevertheless, it was selected as the test crop because, owing to its winter hardiness, it is by far the most common pas- ture species in Finland.The fate of added Na in soil was also investigated by determining cation exchange isotherms for the cation exchange pairs Na/K and Na/Ca.

Material and Methods
The effect of Na on the growth and chemical composition of timothy was studied in a pot ex- periment.The experimental soils (silty clay, loam and organogenic soil.Table 1) were taken from plough layers of cultivated fields in Southern Finland.In the text, the silty clay will be referred to as clay.The high concentrations of Ca and Mg and the relatively high pH of the organogenic soil are probably attributable to liming.
The soils were air-dried and ground to pass a 10-mm sieve.For chemical analyses, part of the soil was ground further to pass a 2-mm or a 0.6-mm sieve (C analysis).The carbon concen- tration was determined using a LECO CHN-900 analyser.The particle size distribution was de- termined by a pipette method.The pH was de- termined in a 0.01 M CaCl 2 suspension at the   solution-to-soil ratio of 2.5:1 (v/v).The electrical conductivity of the soil was determined in a water suspension at the solution-to-soil ratio of 2.5:1.Exchangeable cations were extracted with four successive portions of 1 M ammonium ace- tate, pH 7.0 (Thomas 1982).The soil bulk density was determined for air-dry soil compressed as in the experimental pots.Kick-Brauckmann pots with 7 dm 3 of soil were used in the pot experiment.A small por- tion of soil (0.25 dm 3 ) was taken from each pot to cover the seed and the rest was fertilized.The treatments were: Sodium as Na,SO-10H,O: Potassium as KCI: All the nine combinations were made as 5 replicates for each soil.The pots were fertilized also with other elements as analytical grade chemicals at the following rates (mg dm* 3 ): N as NH 4NO, (150), P as Ca (H,P0 4 )   2 .H,0 (150), Mg as MgS0 4 -7H 2 0 (80), S in sulfates of Mg, Cu, Mn, Fe and Zn (at least 112), Cu as CuS0 4 -5H,0 (4), Mn as MnS0 4 -H,O (4), Fe as (2), Zn as ZnS0 4 -7H,0 (3) and B as H,BO, (2).
The seed of timothy (300 mg Phleum pratense L. cv.Tuukka) was sown on the fertilized soil and covered with unfertilized soil.Three crops of timothy were harvested.The second and the third crop were fertilized with solutions ofN and K at the same rates as at the beginning of the experiment.Sodium was applied only at the be- ginning.The plants were grown outdoors under a glass roof from May to September and watered with deionized water.The first crop was cut 59 days after planting, the second and the third one after 33 days' growth.The plant material was dried at 65°C and analyzed for Ca, Mg, Na and K according to a dry combustion method by Helrich (1990).Potassium and Na were analyzed by flame photometry, Ca and Mg by atomic ab- sorption spectroscopy.A known sample was in-eluded in every analysis series.The coefficients of variation (23 observations) for the analysis of the sample were: Na 17.0%, K 3.6%, Ca 2.9% and Mg 1.1%.
Exchange isotherms were determined for Na/ Ca and Na/K exchange on the three soils by a modified method of Levy et al. (1988).Isotherms were determined using 5 g of soil (2 g of orga- nogenic soil) in duplicates.The soil samples were first equilibrated with 25 ml of a mixture of solutions of NaCl and CaCI 2 or NaCI and KCI.Seven different cation equivalent ratios (0, 15,  30, 50, 70, 85 or 100 % of Na) were used for both isotherms, and the chloride concentration of the first equilibration solution was 0.5 mol 1 1 .I 1 .The suspensions were shaken for 20 min, centrifuged, and the supernatant solution was dis- carded.This was repeated three times, followed by three similar steps using 50 ml of 0.01 M solutions while maintaining the original Na/Ca or Na/K ratios.The last 0.01 M solutions were filtered through Schleicher & Schiill 589 3 (blue ribbon) filter paper and used to determine Na and K by flame photometry and Ca by atomic ab- sorption spectroscopy.The amount of equilibration solution remaining in the soil samples after equilibration was determined by weighing the centrifuge tubes containing the soil and the so- lution.Thereafter the exchangeable cations ad- sorbed on the soil samples were determined by three repeated extractions with 30 ml of 1 M ammonium acetate (20 min shaking, centrifuging and filtering through Schleicher & Schiill 589 3 filter paper).The extracted cations were determined as above.The amount of cations held in the soil by the remaining equilibration solu- tion was deducted from these results.

Results
The dry matter yields were not affected by Na treatments.On the organogenic soil, K fertiliza- tion elevated the yields of the second and third harvest by 10 and 51%, respectively (p<0.001)Table 2. Dry matter yields (g pot') at the three levels of K application.Treatments: K n =o mg dm 3 ,K,=3x100 mg dm' 3 , K 2 =3x200 mg dm 3 .

Clay
Loam Organogenic soil Harvest All K levels All K levels K 0  2).Yet, there were no visible K deficien- cy symptoms in any treatment of the experiment.Without Na and K application (Na 0 K 0 ), the average Na concentration of timothy (Table 3) was 0.24 g kg 1 (range 0.17-0.30g kg '), exclud- ing the plants grown on the clay in the first har- vest that had a Na concentration (0.57 g kg ')  deviating from the other corresponding results.When no Na was applied, K application seemed to elevate the Na concentration but the effect was not statistically significant.
Application of Na increased the Na concen- tration of timothy on all soils and in all harvests (pccO.OOl).The effect was greater for the loam and the organogenic soil than for the clay.The highest Na concentrations of timothy were reached without K and with the highest rate of Na (Na 2 K 0 ) in the third harvest: 6.9 gkg' on the loam and 5.6 g kg ' on the organogenic soil.
Sodium was much weaker than Ca or K in competition for cation exchange sites (Fig. 1).In Na/K exchange, Na was least efficient com- petitor for clay and somewhat more efficient for the two lighter soils.In equimolar equilibrium solution (Na/K), Na occupied 24, 33 and 38% of the exchange sites on clay, loam and organo- genic soils, respectively.Sodium was even a weaker competitor for exchange sites with Ca than with K.In an equilibrium solution contain- ing 50% of both Na and Ca (expressed in mmol of charge dm' 3 ) there was only 5, 4.5 and less Table 3. Sodium and potassium concentrations of timothy in the pot experiment (g kg 1 dry matter).Treatments: Na | =200 mg dm ', Na 2 =400 mg dm ', K,=3xloo mg dm 3 , K 2 =3x200 mg dm 3 .

46'
Each soil and harvest was tested separately.Means with same superscripts do not differat p = 0.05.
than 3% of exchange sites occupied by Na on clay, loam and organogenic soils, respectively.Application of K drastically depressed the effect of Na fertilization (p<0.001) even though added K was more likely to be bound on the cat- ion exchange sites than Na.In the Na/K ex- change, K was least efficient competitor on the organogenic soil, and the effect ofK was strongest on this soil; K, application decreased the Na concentration of timothy at the Na 2 level in the third harvest to one-eighth of that obtained with- out K application (Table 3).When plants were fertilized with K, the targeted Na concentration of 2 g kg 1 was reached only with the higher Na level in the third harvest on the loam.The sodi- um concentration of timothy increased most ef- fectively when the K concentration of the plants decreased to or below 15 g kg ' (Fig. 2).These concentrations, suggesting K deficiency, oc- curred in plants grown without K application in the second and third harvest on the organogenic soil and in the third harvest on the loam.
Application of Na did not decrease the con- centration of K in timothy at any K level.Figure 3 demonstrates the abundance of K compared to other cations in the plants fertilized with the same amounts (mg dm 3 ) of Na and K.When expressed in mol dm 3 , the plants received more Na than K.Despite the similar amounts added as fertilizer, the first harvest of timothy grown on mineral soils took 50 times and timothy grown on the organogenic soil 80 times more K than Na.A tenfold increase in Na concentration from the Na 0 level with an equivalent decrease in K concentration would theoretically lower the K concentration by only about 4 g kg' 1 at the most.
Not even this effect seems to be feasible.
Sodium application decreased the Ca concen- tration of timothy on all soils and for all har- vests (p<0.001).The mean Ca concentrations were 8.5, 7.2 and 6.6 g kg' 1 in the treatments Na 0 , Na ( and Na 2 , respectively.In the third harvest on the organogenic soil, the Ca concentration of timothy was as much as 35% lower in the Na 2 K 0 (10.7 g kg ') than in the Na () K (| (16.5 g kg ') treat- ment.Sodium application did not influence the Mg concentrations which, owing to Mg fertili- zation, were rather high in the experiment.Po- tassium application strongly lowered the Mg concentration of timothy (pccO.OOl).The aver- age Mg concentration was 4.6, 3.1 and 2.6 g kg' 1 in the treatments K O , K, and K 2, respectively.
One of the aims of Na fertilization has been to decrease the excessive concentration of K in herbage and thus to decrease the K/(Ca+Mg) ratio of fodder.In the present study, Na fertili- zation, on the contrary, increased the ratio (p<0.001) because it decreased the Ca concen- tration but did not affect the K and Mg concen- Fig. 1.Cation exchange isotherms for the cation pairs Na/ K and Na/Ca.Sodiumon the exchange sites and in the equilibrium solution expressed as a fraction of total positive charge.
trations (Table 4).However, the K/(Ca+Mg) ra- tios remained below 2.5 in all treatments, and the unfavourable effect ofNa diminished towards the end of the experiment.Without K application (K 0 ) the K/(Ca+Mg) ratios decreased to very low values in the third harvest (range 0.2-0.8).Potassium had a strong enhancing effect on the ratio (p«0.001)throughout the experiment.Sodium fertilization lowered the K/Na ratios in the plants (p«0.001) which is considered to improve the quality of fodder.
In the Na n K 0 pots, 22-34% of the native exchangeable Na was taken up by plants while in the Na Q K 2 pots, 52-71% was taken up.Howev- er, the reserves of exchangeable Na were not de- pleted during the experiment in clay while there was a consistent decrease in the organogenic soil and inconsistent changes in the loam (Table 5).Applied Na elevated the concentration of ex- changeable Na in the soil to very high levels.At the end of the pot experiment, Na extracted from the soils of the highest Na treatment represent- ed 11, 14 and 6% of all extracted cations (mol of charge dm 3 ) in clay, loam and organogenic soils, respectively.The heavy fertilization also elevated the electrical conductivities of the soils.The maximum values at the end of the experiment, measured in the Na 2 K, pots, were 0.8, 0.6 and 1.0 dS nr 1 in clay, loam and organogenic soils, respectively.
In pots fertilized with K, the apparent utili- zation of added Na was less than 4% on all soils and at the K 2 level it remained below 0.8% for the clay and the organogenic soil.At the K 0 level, the utilization of Na was somewhat higher    4. K/(Ca+Mg) ratios (mol of charge kg l ) of timothy in the pot experiment.Treatments: Na,=2oo mg dm \ Na,=4oo mg dm -', K,=3xloo mg dm -', K 2 =3x200 mg dm -\ Clay Loam Organogenic soil Harvest Each soil and harvest was tested separately.Means with same superscripts do not differ at p=0.05.
(8-11%) for the loam and the organogenic soil but for the clay it was below 4% in all treatments.By way of comparison, utilization of added K was 68-75% for the organogenic soil, 34-64% for the loam and 21-45% for the clay.The re- serves of soil K decreased at the K () and K, lev- els for all the soils, for the organogenic soil also at the higher application rate (K 2 ) (Table 5).The Table 5. Exchangeable Na and K concentrations of the experimental soils (mg dm ') before and after the pot experiment.total uptake of K by plants (three harvests) at the K (| level was 325, 199 and 145 mg dm 3 for clay, loam and organogenic soil, of which 120, 89 and 76 mg dm 3 was non-exchangeable K, respectively.

Discussion
The Na concentration of timothy not fertilized with Na was up to 10 times higher in the present pot experiment than generally found in field con- ditions in Finland (Kähäri and Nissinen 1978,  Jansson 1986) even though the experimental soils had an average Na status as compared to the soils of Finland (Sippola and Tares 1978).The high Na concentrations can partly be explained by a high N fertilization rate (450 mg dm' 3 corre- sponding to 900 kg ha ').In field conditions, N fertilization of 600 kg ha 1 has increased the Na concentration of mixed ley by a factor of 6.7 as compared to unfertilized ley (Rinne et al. 1974).Due to the high Na concentrations of timothy in the present pot experiment, the results cannot be applied quantitatively to field conditions, but they give qualitative information about the dif- ferent interactions between cations in the nutri- tion of timothy.In normal cultivation the plants are better supplied with K than in the K 0 pots of the present experiment.Therefore, the response of timothy to Na fertilization in field conditions would most likely be closer to that observed at the K, and K 2 levels and far from the higher re- sponses measured in the K 0 pots.
The increased Na concentration of timothy caused by the application of K at the Na 0 level can be explained by cation exchange.Added K effectively displaces Na from the cation ex- change sites of soil, as was shown in the Na/K exchange studies.A higher Na concentration in the soil solution consequently promotes Na uptake by timothy.However, the observed increase of Na concentration brought about by K application is marginal.The slight increase in plant K concentration upon Na addition in the K 0 treat-ment on loam and organogenic soil reflects the exchange between added Na and soil K.
Calcium is the dominating exchangeable cat- ion in the soil, and the Na/Ca exchange equilibria suggest that added Na remained nearly completely in the soil solution.The particularly high selectivity for Ca of the organogenic soil is probably due to the complexation of Ca with the functional groups of organic matter (Mcßride 1994).In spite of the lowest selectivity for Na over Ca in the organogenic soil, the highest Na concentrations were obtained on the loam.At the end of the experiment the percentage of Na of the sum of cations (in mol of charge dm 3 ) ex- tracted by ammonium acetate was highest in the loam, the Na activity of the soil solution conse- quently being highest in this soil.The difference in Na activity in the soil solution may thus ex- plain the difference in the Na concentration of timothy grown on the mineral and the organogenic soils.The high original Ca concentration of the organogenic soil may also have depressed the uptake of Na through cation antagonism.In K uptake the selectivity by plants seemed to dom- inate the selectivity in cation exchange.The dif- ference in Na concentrations between the mineral soils is due to the larger reserves of K in clay.
Indeed, the most pronounced phenomenon in the present study was the dependence of Na uptake on the K supply, either in the form of native or added K. The effect of Na application on Na concentration of timothy for the three soils (or- ganogenic was at odds with the K supplying power of the soils.The effect of the lower K supplying power of organogenic soils on the Na concentration of timothy was also ev- ident in the study ofKähäri and Nissinen (1978).They reported higher Na concentrations of tim- othy in the province of Lappi, dominated by or- ganogenic and coarse mineral soils, than in the rest of the country.They also observed that tim- othy grown on Sphagnum peat soils contained 122 mg Na kg ' in contrast to the average around 50 mg kg ' in timothy grown on other soils, with- out a marked difference in exchangeable Na in soil.
The increasing efficiency of Na fertilization towards the end of the experiment can, besides a decrease of the reserves of soil K upon K uptake by the plants, partly be caused by gradual filling of storage capacity for Na in roots and conse- quent transport of Na to shoots.Distinctively natrophobic timothy stores as much as 90% of Na taken up in the roots (Jarvis 1982).When this storage capacity is used up, the Na concentra- tion of the shoots starts to increase.It has indeed been observed that the Na concentration of timothy increases when plants age (Jarvis 1982).The same phenomenon, in a weaker form, has been observed with perennial ryegrass, meadow fescue and cocksfoot (Rinne et al. 1974, Smith  et al. 1980).
The effect of Na fertilization on the Ca concentration and K/(Ca+Mg) ratio of timothy was opposite to that observed with perennial ryegrass in field conditions (Chiy and Phillips 1993).However, in Chiy and Phillips' experiment, con- siderably natrophilic ryegrass was grown in sub- optimal K conditions: Na fertilization increased the Na concentration of herbage substantially, increased the yield and Ca concentration and even lowered the K concentration of herbage.The apparent recovery of Na by ryegrass in that field experiment was 70%, which is considera- bly higher than in the present pot experiment.The difference between timothy and ryegrass supplied adequately with K would probably be smaller than these results suggest.

Conclusions
The present study shows that timothy has such a preference for K that even an excessive K con- centration in a plant cannot be suppressed by ample Na additions.The strong reduction of Na uptake by native and added K makes it difficult to elevate the Na concentration of timothy sward by Na applications in practice.As long as K con- centration of timothy is at a level sufficient for maximum growth or higher (luxury consumption), the Na concentration of shoots cannot be effectively increased by Na application.It is possible to elevate the Na concentration of tim- othy substantially only when plants are in K de- ficiency.However, a deficiency of K endangers the production of maximum yield and it may be unreasonable to produce grass high in Na at the expense of the yield.If Na is applied to timothy in field conditions, the utilization rate cannot be expected to be high.Other grass species like cocksfoot and ryegrass utilize Na more effectively (Smith et al. 1980, Jarvis 1982).Cultivation of these less natrophobic species as pure stands or as mixtures with timothy is a more realistic alternative to elevate the Na content of forage than Na fertilization of timothy.

Fig. 2 ,
Fig.2, Relationship of the concen- trations of Na and K of timothy fertilized with two Na levels (Na, = 200 mg dm Na 2 = 400 mg dm 3 ).

Fig. 3 .
Fig. 3. Na, K, Ca and Mg concentrations of timothy (mmol of charge kg ') in the first harvest.The plants received sim- ilar Na and K fertilization (200 mg dm 1 , treatment Na^,).

Table
. Properties of the experimental soils.