Competition and yield performance in mixtures of oats and barley nitrogen fertilization , density and proportion of the components

Competition and yield performance in mixtures of barley and oats were evaluated from addition series experiments (three experiments) in 1983 and in 1984. Three doses of nitrogen fertilization (10 kgN/ha, 40 kgN/ha and 80 kgN/ha) were applied. In the first year the components were Agneta barley and Veli oats and in 1984 in addition to the previous combination also Ida barley and Veli oats were included. The competitive relationship between components was analysed by replacement series model and by regression analysis. The results showed that the dominant component according to the regression analysis was also dominant according to the indices of the replacement series model independently of density and proportion. Barley was generally more competitivethan oats. The dominance of barley usually increased with increasing nitrogen fertilization, especially in the mixture of Agneta and Veli. All the yield components of the barley plants increased with the decreasing proportion of barley in the mixture. In 1983, some mixtures overyielded significantly (p<0.05). The relative yield total being usually greater than one indicated yield advantage. In 1984, oats suffered from insect damage and neither barley cultivar was able to compensate enough so no overyielding occurred. The relative yield total was lower than one and thus no yield advantage was achieved. Index words: Competition, yield advantage, barley, oats, mixtures


INTRODUCTION
interest has been paid to crop mixtures for two main reasons: an increase in yield brought about by the complementary habits of as- yield over locations and seasons due to the ability of at least one genotype in the mixture to yield well in adverse conditions (Taylor   1978).Approximately 50% of the barley and oats grain produced in Ontario is from mixsociated genotypes, and greater stability of tures of the two components (300 000 ha) (Fejer et al. 1982).
In several studies of barley-oats mixtures grown for feed, grain yield increases over the mean of the components in monoculture have been observed and even overyielding has oc- curred (Salminen 1945, Van Dobben 1953,  Bebawi and Naylor 1978, Taylor 1978,   Fejer et al. 1982).Ontario provincial agricultural statistiscs show that mixed grain consis- tently outyielded pure stands as reported by Fejer et al. (1982).In trying to combine the two species so that there was less mutual com- petition at critical stages, Syme and Bremner  (1968) found that mixture yields did not ex- ceed the better component and were usually similar to mid-component.
In most cases of mixtures of oats and bar- ley the yield of the mixture is compared with the average of the yield of the two monocul- tures.A higher yield of the mixture is inter- preted as an argument for mixed cropping.This indicates that the yield advantage of mixtures is not always completely assessed.This is because without calculation of relative yield total the interpretations based on the ratio of actual and expected yields can be misleading, especially in cases where compensation occurs (Willey 1979).Compensation seems also to be the most common situation with mixtures of barley and oats, i.e. the competitive abili- ties of barley and oats differ (for example Salminen 1945, de Wit 1960, Syme and  Bremner 1968, Fejer et al. 1982).
A mixture of species might more efficient- ly utilize the resources and therefore yield more than the pure stand.This may indicate that intraspecific competition is more severe than interspecific interference with growth.For these reasons mixtures might be expected to show a yield advantage (Spitters 1983).
To achieve an accurate assessment of the relative strengths of intra-and interspecific com- petition in mixtures of barley and oats, these experiments were conducted.
In the experiments described here, replacement series (substitutive) (de Wit 1960,   Harper 1977, Connolly 1986) at three total plant densities of barley-oats mixtures and monocultures were used to assess the competitional relationship between species and the yield advantage of mixtures.The design is characterized by the term addition series (Spitters 1983).
Two approaches were used to analyse com- petition.The first approach was to use mea- sures of competitive abilities and combining abilities of varieties based on the relative yield responses according to the de Wit model (de Wit 1960).The other approach is based upon linear regression with the reciprocal of average plant grain yield as the dependent varia- ble and density as the independent variable.The reciprocal yield model was expanded for multiple genotypes by Wright (1981) and Spitters (1983).

MATERIALS AND METHODS
The addition series field experiments were carried out in 1983 (one experiment) and in 1984 (two experiments) at the experimental farm of the University of Helsinki in Helsinki Viikki (60°13'N, 25°00'E) with barley and oats sown separately and in mechanical mixtures.In 1984, the experiments situated side by side.In 1983, the soil was silty clay with pH 5.6 and in 1984 finer fine sand with pH 5.4.
Experimental design and management.A split-split-plot design (nitrogen levels in main plots, total densities in subplots and genotypic composition of stand in subsubplots) was used with three blocks.The subplot size was 10 m 2 (1.25 m x 8 m) with rows spaced 12.5 cm apart.In 1983, the varieties were Agneta bar- ley and Veli oats and in 1984, in addition to previous combination, also Ida barley and Veli oats were included.The general charac- ters of the cultivars are described elsewhere (Jokinen 1991 a).
Sampling and analyses.The number of plants in each plot were determined by count- ing the number of seedlings in four randomly selected 1-m-long rows/plot about three weeks after sowing before the start of tillering.Simi- larly the number of generative shoots in 1983   was determined after the complete ear emer- gence of the cultivars.The height of the stands was estimated visually as well as the emergence time of seedlings.Four weeks after sowing in 1984 samples were taken in three randomly selected 1-m-long rows/plot for determination of the total above ground dry matter of the plants.
From each mixture yield a 50 g sample was taken for determination of the seed yield of the barley and oats components.The separat-ed samples of each mixture as well as samples of each pure stand yield were used for deter- mination of 1000 grain weights (g) (3 x seeds/sample) in 1983.The number of grains/head was calculated using the data of yield, number of generative shoots and grain weight.
Relative yield (RY) and relative yield total (RYT) were calculated according to the meth- od of de Wit and van den Berg (1965).Com- petitive ratio (CR) was determined according to the method of Wiley and Rao (1980).The mean yield/area was calculated before com- puting the indices.Details of the calculations are described elsewhere (Jokinen 1991 b) A discussion of the use of hyperbolic yielddensity equations in various situations has been given elsewhere (Wright 1981, Spitters 1983, Firbank and Watkinson 1985, 1990,  Connolly 1987, Roush et al. 1989).The method used here was described previously (Jokinen 1991 c).
Data on the plant dry weights, the grain yields, 1000 grain weight and the number of generative shoots were subjected to analyses of variance for split-split-plot design (Steel and Torrie 1980).Mean separation was ac- complished by Tukey's honestly significant difference test (HSD) (P = 0.05) (Steel and  Torrie 1980).
Table I.The influence of nitrogen fertilization and proportion of oats in the stand on the phytomass accumulation (dry weight mg/plant) of Veli oats during the first month of growth in 1984 in two barley-oats experiments (Veli/Ida and Veli/Agneta).The analysis of variance is done separately for each experiment.Dry weight means in the average columns and in the average rows followed by the same letter are not significantly different at the 5% level (HSD   test).The first leaves of barley were larger than those of oats (data not given).The number of seedlings in each plot was about the same (0.95 -1.05) as expected (data not given).
The average phytomass of the oats was approximately the same in both experiments (Table 1).Both barley varieties were over twice heavier than the oats (Tables 1 and 2).The average phytomass of all the varieties de- creased with increasing density (data not given).The seedlings of Agneta were heavier than those of Ida.Unlike the oats, the phytomass of the barley varieties increased with decreas- ing proportion of the species in the mixture.Both barley varieties were more competitive than the oats competitive ratio varying from 1.21 to 1.52 (data not given).The relative yield totals varied from 0.96 to 1.05 (data not given).
No lodging occurred in 1983.In 1984, the pure stands of barley at the highest density and the highest level of nitrogen fertilization were the most lodged (Table 3).

Grain yields
In 1983, the mean yield of the experiment of Veli oats and Agneta barley was 5446 kg/ha (Table 4).The analysis of variance showed sig-Table 2. The influence of nitrogen fertilization and proportion of barley in the stand on the phytomass accumulation (dry weight mg/plant) of Ida barley and Agneta barley during the first month of growth in 1984 in two barley-oats experiments (Veli/Ida and Veli/Agneta).The analysis of variance is done separately for each experiment.Dry weight means in the average columns and in the average rows followed by the same letter are not significantly different at the 5%  3. Lodging of the stands (Vo of area) in 1984.(-=nolodging, 100 = completely lodged, Ag = Agneta/Veli, Id = Ida/Veli).
Nitrogen fertilization (kgN/ha) nificant (p<0.05)interaction between the nitrogen fertilization and the proportion of the components, and between the density and the proportion of the components.At the lowest level of nitrogen two out of three mixtures yielded significantly more (approximately 9%) than the pure stands, i.e. the mixtures over- yielded.At the highest density all the mixtures overyielded significantly (approximately 7%).
Comparison between the actual and expected yields of the mixtures (50:50) shows that all the mixtures were more productive than monocultures.
In 1984, the mean yield in both experiments was lower than in the previous year (Tables 5 and 6).This was due to the very low yield of oats because of frit fly {Osdnella frit) dam- age.In general, the grain yield of the stands increased with increasing proportion of bar- ley in the mixture and no overyielding oc- Table 4.The influence of nitrogen fertilization, density and the proportion of barley (Agneta) and oats (Veli) in the mixture on the grain yield (kg/ha) of the stands in 1983.A/E is the ratio of the actual and expected yield of the mixture of 50:50.Grain yield averages within each treatment (nitrogen fertilization, density and proportions) followed by the same letter are not significantly different at the 5% level (HSD test).Comparison between the grain yield averages of different proportions is done at different levels of nitrogen and density (interaction statistically significant).

Proportion Density
Nitrogen fertilization (kgN/ha) curred.It is important to note that in 1984 in both experiments the actual yields of mixtures (50:50) were higher than expected in some cases (Tables 5 and 6).
Relative yields (RY), relative yield totals (RYT) and competitive ratio (CR)   in 1983, the relative yield of Veli oats was higher than expected only at the lowest level of nitrogen fertilization (Fig. 1).In 1984, the relative yield of oats was always lower than expected (Figs. 3 and 5).In both years the rela- tive yields of barley were usually higher than expected (Figs. 1, 2 and 3).
In general, barley was more competitive than oats (CR> 1) (Figs. 2, 4 and 6).Only in 1983 at the lowest level of nitrogen fertiliza- tion were oats as competitive as barley in some cases (Fig. 2).In 1984, Agneta was more corn- Table 5.The influence of nitrogen fertilization, density and the proportion of barley (Agneta) and oats (Veli) in the mixture on the grain yield (kg/ha) of the stands in 1984.A/E is the ratio of the actual and expected yield of the mixture of 50:50.Grain yield averages within each treatment (nitrogen fertilization, density and proportions) followed by the same letter are not significantly different at the 5% level (HSD test).Comparison between the grain yield means of different proportions is done at different levels of nitrogen and density (interaction statistically sig- nificant).Figure I.The influence of density (plants/m ! ), nitrogen fertilization (kg N/ha) and proportion of the components on the relative yields (RY) of Agneta barley and Veli oats, and on the relative yield totals (RYT) of the mixtures in 1983.
Figure 2. The influence of density (plants/m ; ), nitrogen fertilization (kg N/ha) and proportion of barley on the competitive ratio (OR) of Agneta barley over Veli oats in 1983.
Figure 3.The influence of density (plants/m 2 ), nitrogen fertilization (kg N/ha) and proportion of the components on the relative yields (RY) of Agneta barley and Veli oats, and on the relative yield totals (RYT) of the mixtures in 1984.petitive over oats than was Ida.In 1984, Agneta was more competitive than in the previous year.Especially Agneta barley was the most competitive at the highest level of nitrogen fer- tilization.In 1984, unlike the previous year, the competitive ratio of Agneta usually increased with increasing proportion of barley.As a rule, the relative yield totals exceeded one in 1983 (Fig. 1).In 1984, the relative yield totals of both mixtures were close to or lower than one (Figs 3 and 5).

Regression models
The regression equations accounted for 90-96% of the variation in grain yield of both species (R 2 = 0.90-0.99)(Tables 7, 8 and 9).Only in 1984 in the mixture of Agneta barley and Veli oats were the regression coefficients of oats not statistically significant in regression equations for oats.
As a rule, the intraspecific competition of barley was more severe than the interspecific competition and vice versa for oats.Barley Table 6.The influence of nitrogen fertilization, density and the proportion of barley (Ida) and oats (Veli) in the mixture on the grain yield (kg/ha) of the stands in 1984.A/E is the ratio of the actual and expected yield of the mixture of 50:50.Grain yield averages within each treatment (nitrogen fertilization, density and proportions) followed by the same letter are not significantly different at the 5% level (HSD test).Figure 5.The influence of density (plants/m ! ), nitrogen fertilization (kg N/ha) and proportion of the components on the relative yields (RY) of Ida barley and Veli oats, and on the relative yield totals (RYT) of the mixtures in 1984.
Figure 6.The influence of density (plants/m 1 ), nitrogen fertilization (kg N/ha) and proportion of barley on the competitive ratio (CR) of Ida-barley over Veli-oats in 1984. .NDI (Niche differentiation index) = (Bbb/Bba)/(Bab/Baa). 1/W is the reciprocal yield of an individual plant (grain yield/plant).B 0 is the reciprocal of the theoretical maximum yield of an individual, B 1 describes influences of intragenotypic competition, B 2 describes influences of intergenotypic competition, N is plant density and RC predicts relative competitive ability of each genotype.p< 0.001 for B 1 and B 2 in each model.
Table 8.Multispecies reciprocal yield models (1/W = Bo+ BINI + 82N2) for interactions between barley (Agneta) and oats (Veli) grown at three levels of nitrogen fertilization in b-values x lO-3 .NDI (Niche differentiation index) = (Bbb/Bba)/(Bab/Baa).1/W is the reciprocal yield of an individual plant (grain yield/plant).B 0 is the reciprocal of the theoretical maximum yield of an individual, B 1 describes influences of intragenotypic competition, B 2 describes influences of intergenotypic competition, N is plant density and RC predicts relative competitive ability of each genotype.p<o.ool for B 1 and B 2 in each model of barley and for B 2 in each model of oats.B 1 in the models of oats is not significant.
Table 9. Multispecies reciprocal yield models (1/W = Bo+ BINI + 82N2) for interactions between barley (Ida) and oats (Veli) grown at three levels of nitrogen fertilization in  benefitted at the expense of oats.The exceptional case when the intraspecific competition in the mixture was stronger than interspecific competition for both species (81/B2> 1)was at the lowest level of nitrogen fertilization in 1983 (Table 7).Then both components benefitted from mixed culture.In this case the asymptotic yields of both components grown in mixture (1/(B1 -I-B 2)) were higher than the asymptotic yields of both components grown in monoculture (1/B1).Barley was a stronger competitor than oats as determined for the ratio of regression coefficients from barley (RC = 81/B2).Agneta was a stronger competitor in 1984 than in 1983.In 1984, Agneta was more competitive against oats than was Ida.In most cases the relative competitive ability of barley increased with increasing nitrogen fertilization.
Only in 1983 was the overall intraspecific competition greater than the overall inter- specific competition (NDI > 1) independent of nitrogen fertilization.In 1984 the competition was more severe in the mixture of Agneta and Veli than in the mixture of Ida and Veli (NDI Agneta/Veli < NDI Ida/Veli).
The square root of the product of the interspecific competition coefficients [(Babxßbaf'/i] was less than the intraspecific competition coefficients of barley and oats only in two cases in 1983.In these situations a mix- ture of optimum proportions will yield more than both monocultures, i.e. a mixture will overyield.

Yield components
In 1983, the addition of nitrogen, change Table 10.The influence of nitrogen fertilization, density and the proportion of barley (Agneta) and oats (Veli) in the mixture on the number of generative shoots per plant of Agneta barley in 1983.Shoot number averages within each treatment (nitrogen fertilization, density and proportions) followed by the same letter are not significantly different at the 5% level (HSD test).Comparison between the shoot number means of different proportions is done at different levels of density (interaction statistically significant).

Proportion Density
Nitrogen fertilization (kgN/ha) of the total density of stands or the growth in a mixture as compared with pure culture affected all the yield components (the num- ber of generative shoots per plant, the 1000 grain weight and number of grains per head) of both species in certain extent (Tables 10-15).

Advantages of mixtures
The results of the present experiment suggest that overyielding may occur in the mixtures of oats and barley under certain conditions.Other studies of mixtures of barley and oats suggest that the yield of a mixture can be above that of the better component (Salminen 1945, van Dobben 1953, Bebawi  and Naylor 1978, Taylor 1978, Jokinen 1991 a).Syme and Bremner (1968) and Fejer  et al. (1982) found that mixture yields did not exceed those of the better component.
In addition to overyielding, mixtures can be advantageous over monocultures if the yield of the mixture exceeds the mid-component but are not necessarily so.For example in 1984 the actual yields of mixtures exceeded the expected in many cases, however, theresults of rela- tive yield total of a given mixture indicated no yield advantage.Thus when relative yield to- tal did not exceed one the same yield of bar- ley and oats might have been obtained with monocultures as with mixtures, without changing the total area of land (Willey 1979).At least from field experiments the rela- tive yield total could be assessed for the proper Table 11.The influence of nitrogen fertilization, density and the proportion of barley (Agneta) and oats (Veli) in the mixture on the number of generative shoots per plant of Veli oats in 1983.Shoot number averages within each treatment (nitrogen fertilization, density and proportions) followed by the same letter are not significantly different at the 5% level (HSD test).Comparison between the shoot number means of different proportions is done at differ- ent levels of density (interaction statistically significant).

Proportion Density
Nitrogen fertilization (kgN/ha) There are only a few published experimental results with barley-oats mixtures from which it is possible to calculate relative yield totals.The relative yield total of the mixture of oats and barley was close to one (1.03) as calculated by the author from the results of eight experiments conducted by Salminen (1945).This indicates no or a very slight yield advantage.The results calculated by de Wit   (1960) indicated that in general the yield of barley or oats is proportional to the relative space occupied by these crops.The relative yield totals calculated by the author from the experiments of Syme and Bremner (1968)  varied from 0.90 to 1.15 with two out of sev- en values being lower than one.The calculat- ed relative yield totals (1.07, 1.15) from the two experiments of Fejer et al. (1982) as well as the results of the present experiment in 1983 suggest that cropping of mixtures of barley and oats may be of benefit.However, more experiments in different environments are needed to provide support for the practical use of mixtures.
In addition to possible yield advantages there are other benefits of growing barley and oats in mixture such as prevention of lodging (de Wit 1960).The results of the present ex- periments in 1984 suggest that lodging may be reduced by growing mixtures.The decreased lodging of mixtures compared with monocul- tures may be because of shorter barley plants in mixtures (K.J. Jokinen unpubl.).The in- creasing light intensity during the growth of barley plants is known at first to increase and Table 12.The influence of nitrogen fertilization, density and the proportion of barley (Agneta) and oats (Veli) in the mixture on thousand grain weight (g) of Agneta barley in 1983.Grain weight averages within each treatment (nitrogen fertilization, density and proportions) followed by the same letter are not significantly different at the 5% level (HSD test).Comparison between grain weight means of different proportions is done at different levels ofnitrogen fertilization (interaction statistically significant).then to reduce plant height (Briggs 1978 p.274).Thus in the monoculture of barley, plants might shade each other more than in mixed stands light being possibly a limiting factor especially at high levels of nitrogen fertilization and at high densities.
Compensation in 1984, the dominant-suppression relation- ship between barley and oats was not always complete (RYT < 1).This indicates that in mixtures barley interfered with the yield for- mation of damaged oats more than expected without benefitting by itself.Thus barley was not flexible enough especially at low densities.According to de Wit (1960), the ability of un- damaged component to compensate depends on the time of damage.Thus the compensa-tion relates to the determination of the yield components and flexibility of the plants dur- ing the course of the development as well as the total density of the stands.One has to notice that in an extreme case the competition experiment can degenerate into a spacing ex- periment for one component.

Competition models
Although there were no profound dis- crepancies between the results of the two different approaches for analysing competitive interactions between components, the regression approach provided a more flexible framework for mixture studies than the conventionalreplacement analysis.The regression analysis uses a model of competition that al- lows the yields of both species in a binary mix-Table 13.The influence of nitrogen fertilization, density and the proportion of barley (Agneta) and oats (Veli) in the mixture on thousand grain weight (g) of Veli oats in 1983.Grain weight averages within each treatment (nitrogen fertilization, density and proportions) followed by the same letter are not significantly different at the 5% level (HSD   test).This is because the shoot growth of species adapted to high nutrient conditions (with intense shoot competition) usually responds more to increased nutrient supplies than does that of species adapted to low nutrient condi- tions (with intense root competition).
It is important to note that in the 1984study the results did not show thoroughly the effect of the nitrogen gradient on competition.This was because of the weak response of the stands to the added nitrogen possibly due to the high release of nitrogen from the soil.The yields of oats were also depressed due to in- sect damage (frit fly).
The results especially in 1983 demonstrat- ed the effect of nitrogen fertilization on the structure of mixed plant community.There will be a point along the gradient, where Veli oats and Agneta barley can stably coexist by producing the same absolute yields (either number of seeds/plant or weight/plant).From theoretical considerations it follows that a necessary condition for a stable coexistence is that the species populations are regulated in different ways (Braakhekke 1980).Thus bar- ley and oats might be able to coexist because they are not limited by the same resources (differentiation of their realized niches).This means that in mixtures intraspecific competition of both species is greater than interspecif- ic competition which occurred at the lowest level of nitrogen fertilization in 1983.If, how- ever, there is no such differentiation of their realized niches, or if it is precluded by the hab- itat (high nitrogen in 1983 and in 1984), then one competing species (in this case barley) will eliminate or exclude the other (in this case oats).Exclusion occurs when the realized niche of the superior competitor fills those parts of the inferior competitor's fundamen- tal niche provided by the habitat, and the weak interspecific competitor lacks a realized niche when in competition with the stronger competitor (Begon et al. 1986 p. 258-260).
Although the overall intraspecific competition is greater than the interspecific competition in mixtures (NDI> 1), it does not neces- sarily mean that the species can stably coexist in that environment.This is because the strong interspecific competitor will gradually out- compete the weak interspecific competitor as the regression model predicts (in 1983 high nitrogen).Thus NDI when greater than one did not always express that the realized niches of both species differ from each other.The fundamental question is whether NDI is a rele- vant index for evaluating niche differentiation from a terminological point of view.Only when interspecific competition for both species is less significant than intraspecific com- petition the species coexist (81/B2>1), and coexisting competititors may then exhibit differentiation of realized niches.

Yield components
All the yield components of barley tended to associate positively with higher yields per plant obtained from mixtures.The higher grain weight of oats in mixtures did not al- ways compensate for a lower number of ears per plant and lower number of grains per ear, leading usually to a lower yield per plant of oats in mixtures.The results for yield com- ponents suggest that the type of the plants (individuals of the same species or different spe- cies) and their relative amount in the neighbourhood are also significant in respect to the formation of different yield components and not only the total density.As a rule, the yield formation of the individual plant seems to be a rather complex phenomenon in different microenvironments.
the barley seedlings emerged first, about three days earlier than the oats.

Figure 4 .
Figure 4.The influence of density (plants/m 2 ), nitrogen fertilization (kg N/ha) and proportion of barley on the competitive ratio (CR) of Agneta barley over Veli oats in 1984.
*b-values x 10~3 .NDI (Niche differentiation index) = (Bbb/Bba)/(Bab/Baa).1/W is the reciprocal yield of an individual plant (grain yield/plant).BO is the reciprocal of the theoretical maximum yield of an individual, B 1 describes influences of intragenotypic competition, B 2 describes influences of intergenotypic competition, N is plant density and RC predicts relative competitive ability of each genotype.p<0.05 for B 1 and B 2 in each model.
reversals of dominance.