The effect of chloride and nitrogen on nitrate accumulation and yield in beetroot ( Beta vulgaris var . conditiva )

A pot and a field experiment were conducted to evaluate the effects of different nitrogen and chloride fertilizer levels on the nitrate content ofbeetroot. The yield and dry matter content were also determined. Sulphate fertilization was used as a control to chloride fertilization. There was a considerable decrease in the nitrate content of beetroots during the growing season. High nitrogen fertilization caused nitrate accumulation in both experiments. Chloride had a significant decreasing effect on the nitrate accumulation towards the middle of the growing period in the pot experiment. In the field experiment, chloride also decreased nitrate accumulation towards the middle of the growing period, soon after additional application of ammonium nitrate limestone (13.8 % NH4-N; 13.7 % NOj-N). Chloride tends to decrease nitrate accumulation only at an early stage of root development when nitrate is not the only source of nitrogen in the soil. The yield was higheron high nitrogen supply, in the pot experiment also on chloride application. Nitrogen decreased the dry matter content, but chloride had this effect only in the field experiment.


Introduction
High nitrate levels in human food are undesirable, because nitrate may be converted into nitrite which causes methemoglobinemia or is converted into carcinogenic nitrosoamines.Vegetables are the main source of nitrate in the Finnish diet (Penttilä  et al. 1990).
Chloride in the soil is antagonistic to nitrate uptake (James et al. 1970).According to Hähndel  and Wehrmann (1986 b), reduction of the nitrogen supply and addition ofchloride decreases consider- ably the nitrate contents of spinach and lettuce.Van der Boon et al. (1988) reported that a high level of chloride decreases the nitrate content oflettuce and suggested that, in the presence ofammonium nitro- gen, chloride may replace nitrate as a vacuolar osmoticum.Chloride may also replace nitrate as a counter-anion ofcations (Allen and Smith 1986).
The experiments reported here were conducted to find out, whether chloride has the same effect on the nitrate content in beetroot as in lettuce or spinach.The pot experiment was carried out during the growing season of 1989 on an organic soil and the field experiment during the growing season of 1990 on a clay loam soil.

Material and methods
The experiments were carried out in Jokioinen (60°49'N; 23°28'E).The experimental soils were analyzed for potassium, phosphorus, calcium and magnesium extractable in acid (pH 4.65) ammonium acetate (Vuorinen andMäkinen 1955, Kurki et al. 1965), as well as pH and electrical con- ductivity (EC) in water suspension.The soil EC was determined also after the growing season.The boron content was determined by the azomethine- H method (Sippola and Erviö 1977), the organic carbon content by a Leco-analyzer at 1370 °C (Sip- pola 1982) and the particle size distributionby the method of Elonen (1971).The characteristics of experimental soils are presented in Table 1.

225
).A total of twenty treatments were compared (Table 2).The experiment was made with five replicates, of which one was used for sampling 80 days after sowing.Nitrogen was added as NITNC" .Treatments with NaCl and 4 3 Na,S0 4 contained equal amounts of Na.Basic fertilization was given as powdered PKfertilizer (P 7.0 %, K 16.6 %) 10 g per pot.Other nutrients of this fertilizer were: N 2.0 %, Ca 5.4 %, S 12.0 %, Mg 2.5 %, Na 0.5 %, Fe 0.1 %, Cl 0.7 %, B 0.15 %, Cu 0.1 %, Mn 0.7 %, Zn 0.1 %, Mo 0.01 % and Se 0.0016 %.The additional boron (10 mg per pot) was added as boric acid.The soil was limed with CaCO, at the rate of 18 g per pot because of low pH value.The liming material as well as the fertilizing powder and solutions were thoroughly mixed in the soil.
Four kilograms of the fertilized soils (moisture around 55 % on bulk basis) were put into 6 1 plastic pots and compacted slightly.Twenty seeds per pot were placed on the surface and covered with a 2 cm layer of fertilized soil.The beetroot cultivar was 'Little Ball SG', which is a common beetroot culti- var in Finland.A week after emergence the plants were thinned to 12 plants per pot.The plants were watered with de-ionized water once or twice a day.The percolated water was collected and reused.
The first root and shoot samples were collected 54 days after sowing when the plants in the whole experiment were thinned to 6 plants per pot.The second sampling was carried out 80 days after sow- ing by harvesting one replicate.The other four re- plicates were harvested 96 days after sowing.The shoot and root yields were measured.The roots were grated and frozen at -20 °C for laboratory analyses.The nitrate nitrogen content was ana- lysed with a nitrate electrode (ORION 1983, Aura  1985).For the determination of dry matter content, samples were dried at 60 °C for 48 hours.
The results were analysed using the analysis of variance of split-plot design, and Tukey's FISD (Honestly Significant Difference) test was used to determine the significances (p=0.05) of differences between group means (Steel and Torrie 1981).

Field experiment
The field experiment was set up according to a split-plot design, where the main plot was nitrogen fertilization (N |0|) , N 20 0) and the subplot was K 2 S0 4 or KCI fertilization S gg , Cl 95 , Cl |90 ).The treat- ments are presented in Table 2.The experiment was made with four replicates.Nitrogen fertiliza- tion was applied as ammonium nitrate limestone (13.8 % NH 4 -N; 13.7 % N0 3 -N), 70 % in spring and 30 % ten weeks after sowing.In K 2 S0 4 and KCI fertilizations equal amounts of K were given.
Because of the good nutrient status of the soil (Table 1) the only basic fertilization given was 200 kg superphosphate (20 % P).Potassium chloride and potassium sulphate were applied with Tupla-Tume fertilizer drill.Superphosphate and ammonium nitrate limestone were broadcasted by a manually used fertilizer spreader working on the principle of an ordinary fertilizer drill.The fertil- izers were harrowed into the depth of 5 cm.The tractor drove along traffic lanes to avoid soil com- paction.The distance of the tractor wheels was 2 m, which was also the width of the subplots.
Seeds were sown 3 cm deep at a row distance of 40 cm by a Nibex sowing machine.The plants were later thinned to 25 plants/m.There were four rows in a plot, 10 m each.Phenmedipham was sprayed with weed control and dimethoate was applied three times against tarnished plant bug (Lygus rugulipennis).
Topsoil (20 cm) samples were taken 11, 66, 77, 91, 111 and 146 days after sowing from two replicates.The first and the second samples were taken from between the middle and edge rows, and the other samplings from places, where beetroots were grown for plant samples.Soil mineral nitro- gen was extracted with 2 M KCI (Keeney and Nel- son 1982) and analyzed with a Skalar autoanalyzer (Krom 1980, Greenberg et al. 1980).Root samples were taken 75, 82, 89, 97, 102 and 117 days after sowing to determine the nitrate con- tent of roots.The samples were taken from the edges of the middle rows along a length of one meter.Sampling was always done between 8.00 and 10.00 a.m. because of the diurnal change in the nitrate content.Samples were prepared and nitrate contents determined like in the pot experiment.
The root and shoot yields were harvested and measured 102 days after sowing along a length of 4 m from two middle rows (= 8 m).The beetroots were classified according to root diameter into three classes: <4 cm, 4-8 cm and >8 cm.There were no roots larger than 8 cm in diameter.Some beetroots were left in the edges of the plots for determinationof nitrate content later in autumn.
Statistical analysis was made using multivariate analysis of variance to test differences between treatments in all six plant samplings (Littell et al.  1991).The analysis of variance and Tukey's HSD test were used to test significances of differences between group means of samplings (Steel and  Torrie 1981).

Pot experiment
On average, the nitrate contents remained low in the pot experiment (Table 3).Towards the middle of the growing period, 54 days after sowing, the nitrate contents were about ten times higher than at harvest time, the nitrate values increasing with increasing nitrogen levels.
The effect of chloride application on the nitrate contents was not so clear as the effect of nitrogen.
Towards the middle of the growing period, chloride decreased the nitrate accumulation significantly (p=0.05)only in the treatment of maximum nitro- gen and chloride levels.In these samples, the nitrate contents were 2200-2800 ppm in FW below the highest chloride level, but 1900 ppm at the highest chloride level.Two weeks before harvesting and at harvest the nitrate contents were very low.Consequently, the nitrate content was not significantly dependent even on the nitrogen fertilization.However, in the treatment of additional nitrogen fertilization 54 days after sowing, chloride application still decreased the nitrate concentrations.The differences were significant (p=0.05) between the groups of highest chloride level and lowest sulphate level.
Towards the middle of the growing period, at nitrogen levels of 200 and 400 mg/kg soil, more nitrate was accumulated at the highest sodium level.The effect of Na 2 S0 4 can be explained by an increase in soil EC (Table 3) caused by this treatment.
Even though the seedlings emerged slowly in the pots given high rates of NaCl, both root and shoot yields were the highest on the highest chloride supply.In shoot yield the difference between low Na 2 S0 4 and high NaCl levels was significant (p=0.05).Nitrogen application decreased the dry matter content of roots significantly (p=0.05),but chloride had no effect on the dry matter content.

Field experiment
Nitrate nitrogen in soil decreased continuously during the growing season (Fig. 1).The difference in soil nitrate contents between two nitrogen treat- ments was almost double until October.The soil ammonium nitrogen level was low 66 days after sowing.Nitrogen application 67 days after sowing doubled the amount of soil ammonium after two weeks.Then it dropped to a constant low level in two weeks.
The nitrate contents of roots decreased significantly (p=0.01)during the period between 75 and 117 days after sowing.At the nitrogen level of 200 kg/ha, the fresh roots had on average 900 ppm higher nitrate contents than in the nitrogen treatment of 100 kg/ha (Fig. 2).The difference was statistically significant (p=0.01).
There was no significant difference in nitrate accumulation between KCI and applications according to the multivariate analysis of variance (Fig. 3).When every sampling was compared sepa- rately using the analysis of variance, there was a significant difference in the first sampling, 75 days after sowing.The fertilization of S0 4 -S 88 kg/ha caused about 400 ppm higher nitrate content in fresh roots than the chloride treatments.
As compared with the nitrogen fertilization of 100 kg/ha, the nitrogen level of 200 kg/ha had a significant (p=0.01)positive influence on the root and shoot yield.The application ofKCI caused higher shoot yield but no higher root yield compared with K 2 S0 4 fertilization (Table 4).
According to the multivariate analysis of vari- ance, the nitrogen fertilization of 200 kg/ha had a significant (p=0.01)decreasing influence on the dry matter content ofroots (Table 5).The fertiliza- tion of KCI decreased the dry matter content of roots significantly (p=0.01)compared with K 2 S0 4 fertilization.The dry matter contents decreased 89 days after sowing, because ofheavy rainfalls after a dry period of two weeks.Table 4. Effect of different fertilizing treatments on root and shoot yield ofbeetroot in the field experiment.

Discussion
The nitrate contents decreased both in the pot and in the field experiments during the growing season.This decrease has been noticed also by Peck et al.  (1974) and Kallio et al. (1980).
Nitrogen fertilizers increase the nitrate content and yield of beetroot (Peck et al. 1971).This has been noticed also in Finland (Aura 1985), where the recommended nitrogen fertilization level for optimal beetroot yield has been 60-120 kg/ha (Lehtinen 1984, Vuorinen and Takala 1987).Nitro- gen levels over 120 kg/ha impair the quality by increasing the nitrate content.In some experiments even the yield has decreased (Lehtinen 1984).However, differences between growing seasons cause a great deal of variation in beetroot yields, and it is difficult to estimate the nitrogen requirement for optimal yield production.Nitrogen decreased the dry matter contents from 0.5 to 1.0 percentage units, which is in agreement with the results of Aura (1985).Chloride decreased the dry matter contents only in the field experiment.
In the pot experiment, high levels of both nitro- gen and chloride were needed in order to decrease nitrate accumulation.In the final yield no differ- ences were observed as the beetroots used all nitrogen because of good growing conditions.
In the field experiment, chloride decreased the nitrate content only in the first sampling, 75 days after sowing, which was 8 days after additional application of ammonium nitrate limestone.Thus, it may be assumed that chloride needs ammoniumnitrogen in order to decrease the nitrate accumula- tion in beetroot.A similar conclusion was drawn by van derBoon et al. (1988) concerning lettuce.They used nutrient solutions, where it is easy to control the ion concentrations.In field conditions it is very difficult to maintain sufficiently high ammonium levels in soil because of nitrification.
Nitrification inhibitors together with ammonium and urea fertilizers tend to decrease the nitrate accumulation, but high rates ofnitrogen application involve a risk of ammonium toxicity (Goh and Vit-  yakon 1983, Hähndel and Wehrmann 1986 a).   Kallio et al. (1980) found that nitrapyrin caused a 30 % reduction of nitrate content.In their field experiment, no differences were found in the nitrate content of yield between KCI and K,S0 4 applications, when the amount of potassium was 150 kg/ha.According to Cantliffe and Goodwin  (1974), chloride containing fertilizers reduced the nitrate concentration at harvest time in the beetroot petioles, but not in the roots or the blades.

Fig
Fig, I. Soil mineral nitrogen contents in a field experiment during the growing season of 1990.

Fig. 2 .
Fig.2.Nitrate contents in roots ofbeetroot during the grow- ing season of 1990 at two rates of nitrogen fertilization.

Fig. 3 .
Fig. 3. Nitrate contents in roots ofbeetroot during the grow- ing season of 1990 at different rates of chloride and sulphate fertilization.

Table I .
Characteristics of experimental soils

Table 2 .
Fertilization treatments in the pot and field experi-

Table 3 .
Effect of different fertilization treatments on soil EC, and nitrate content, yield and dry mattercontent ofbeetroot in the pot experiment.

Table 5 .
Effect of different fertilization treatments on dry mattercontent ofbeetroot roots in the field experiment.Main and sub plot means without a common superscript letter are significantly different (p<0.05).Means without superscript letters have no significant differences.