The effect of pea cultivar and water stress on root and shoot competition between vegetative plants of maize and pea

Authors

  • Tzehaye Semere,

    1. Department of Agricultural Botany, School of Plant Sciences, University of Reading, 2 Earley Gate, Reading, Berkshire RG6 6AU, UK
    Search for more papers by this author
    • *

      Present address and correspondence: ADAS Rosemaund, Preston Wynne, Hereford, Herefordshire HR1 3PG, UK (fax 01432 820121; e-mail Tzehaye.Semere@adas.co.uk).

  • Robert J. Froud-Williams

    1. Department of Agricultural Botany, School of Plant Sciences, University of Reading, 2 Earley Gate, Reading, Berkshire RG6 6AU, UK
    Search for more papers by this author

Summary

  • 1Improvements in intercrop yields may be achieved through an understanding of yield advantages due to above-ground or below-ground interactions.
  • 2Forage maize and two morphologically contrasting cultivars of pea (leafy cv. Bohatyr and semi-leafless cv. Grafila) were grown alone and in additive mixtures, under two contrasting levels of soil moisture (± water stress).
  • 3The mechanism of competition between maize and pea was studied by separating the effects of root competition and shoot competition, using soil and aerial partitions. Plants were grown in rectangular tanks in a glasshouse.
  • 4Leafy pea cv. Bohatyr was as competitive as maize, both below-ground and above-ground, whereas semi-leafless pea cv. Grafila was less competitive than maize or pea cv. Bohatyr. The greater competitive ability of the leafy pea, both above- and below-ground, was probably due to its greater growth rate, associated with its greater leaf area.
  • 5The competitive ability of maize, relative to peas, was considerably reduced by water stress. Both the root and shoot competitive abilities of pea were greater under water stress, compared with those of maize.
  • 6Relative yield total (RYT) values were significantly greater when maize and pea were subjected to shoot competition only (RYT = 1·76) than when subjected to root competition (RYT = 1·17) or when subjected to both shoot and root competition (RYT = 1·13). This reflects the fact that the effects of root competition were greater than those of shoot competition.
  • 7Root competition decreased the shoot dry weights, plant height and leaf area of both maize and pea, whereas shoot competition had no significant effect on these attributes, indicating that soil resources, i.e. mineral nutrients and water, were more limiting than light.

Introduction

Yield advantage of crop mixtures occurs when the component crops are in complementarity with each other, resulting in more effective use of environmental resources, i.e. light, water and nutrients, compared with when grown alone. Understanding the mechanism is important in understanding yield advantages associated with intercropping.

Physically separating the root and shoot systems of component crops enables the effects of competition for light to be separated from the effects of competition for soil nutrients and water. The separation technique was first achieved by Donald (1958) in his study on Lolium perenne and Phalaris tuberosa (= P. aquatica), and modifications of this technique have been used by many researchers (Aspinall 1960; Snaydon 1971; Barrett & Campbell 1973; Haugland & Froud-Williams 1999) to study the nature of root and shoot interactions between crops and weeds, between component crops in intercropping, and between species of pasture plants.

A detailed review of shoot and root competition was made by Wilson (1988) and Snaydon (1996). Most previous studies on comparisons of root and shoot competition dealt with crop and weed interactions, but only limited studies have investigated intercrops of cereals with legumes (Willey & Reddy 1981; Martin & Snaydon 1982; Martin & Field 1987; Bulson 1991; Tofinga, Paolini & Snaydon 1993).

The competitive ability and resource complementarity between maize and pea was studied using two pea cultivars of contrasting leaf morphology at two levels of soil moisture. Morphological differences have been observed to influence the performance of pea cultivars both in monoculture (King 1978; Pyke & Hedley 1985; Wall, Friesen & Bhati 1991; Knott 1994) and with intercropping (Tofinga, Paolini & Snaydon 1993). Because maize and pea differ in their soil moisture requirements, and have been reported to compete for water in field experiments (Semere 1998), soil water supply was varied in this study and the effects on the yield and competitive ability of the component crops measured.

The objectives of this experiment were to investigate: (i) the nature of competition between maize and pea in mixture; (ii) the possible changes in the relative root and shoot competitive abilities of two cultivars of pea when intercropped with maize; and (iii) the effects of water stress on the competitive effects and resource complementarity between the component crops.

Materials and methods

Experimental design and treatments

The experiment was carried out at The University of Reading, Reading, UK, in a glasshouse maintained at 25 °C day/night temperature with a 16-h day supplementary lighting. Pots were filled with 80% sandy loam soil and 20% peat-moss. Maize cv. Energy (Sharpes International Seeds Ltd., Sleaford, UK), and pea cvs Bohatyr and Grafila (Nickerson Seeds Ltd., Bury St. Edmunds, UK) were evaluated.

The experimental design was a 24 factorial with four replications in a completely randomized block layout. The treatments were as follows:

  • two levels of root competition (± root competition);
  • two levels of shoot competition (± shoot competition);
  • two pea cultivars (a conventional leafy pea cv. Bohatyr, and a semi-leafless cv. Grafila);
  • two watering regimes (± water stress).

The experimental technique used for separating root competition from shoot competition was a modification of that reported by Donald (1958) and similar to that of Schreiber (1967) as plants were grown in alternate rows. Root and shoot competition were separated using soil and aerial partitions producing four combinations of root and shoot competition as follows (see also Fig. 1):

Figure 1.

The planting arrangements of maize (x) and pea (o) showing: (a) full competition; (b) root competition only; (c) shoot competition only; and (d) no competition. Double solid lines represent soil partition (c and d) and broken lines represent aerial partition (b and d).

  • full competition, in which both root and shoot systems were allowed to interact;
  • root competition only, in which the root systems were allowed to interact but the shoot systems were separated above the soil surface by ‘sisalcraft’ reflective aluminium sheet (British Sisalcraft Ltd., Bristol, UK);
  • shoot competition only, in which the shoot systems were allowed to interact but the root systems were separated by growing them in separate pots joined to one another;
  • no competition (i.e. the monoculture treatment), in which both root and shoot systems were separated as described above.

In both shoot competition only and no competition treatments, the aerial partitions were elevated in line with the growth of the tallest plant to ensure that shoot systems were always separated. It should be noted that the effectiveness of an intercrop is measured using the land equivalent ratio (LER) which is based on the relative land requirement for intercrops compared with monocultures. Thus, the land occupied by the intercrop is effectively half of that occupied by the component species altogether. Relative yield total (RYT) is identical to LER and was used in this study.

All treatments were additive series, i.e. the plant density of each component species in any mixture was the same as that in the equivalent monoculture. An additive design was chosen because it allows interspecific competition to be studied separately without being confounded by the effects of intraspecific competition (Jolliffe, Minjas & Runeckles 1984; Connolly 1986; Snaydon 1991, 1996). Thus, in order not to confound the competitive effects of one crop on the other with changes in total density, the two crop species occupied the same amount of space when in competition or when not in competition.

The watering regimes were: (i) 15% of field capacity (i.e. water stress treatment) and (ii) 70% of field capacity (i.e. no water stress treatment). The field capacity of the soil was determined by flooding six small pots of soil, then leaving them to drain overnight; the wet soil was then weighed, and reweighed after being oven dried at 80 °C for 72 h. The volume of water at field capacity was then calculated, and used as a basis for determining the volume of the water required to give 70% and 15% field capacity. Thereafter, a Thetaprobe soil moisture sensor (Delta-T devices, Cambridge, UK) was used to maintain the volumetric soil moisture content in the top 10 cm of soil, starting 17 days after sowing.

Individual boxes were 45 cm long, 31 cm wide and 31 cm deep. Ten maize and 20 pea seeds were planted in each pot in rows. Two weeks after sowing, seedlings were thinned to five maize and 10 pea plants per box; these plant densities were equivalent to 24 plants m−2 for maize and 65 plants m−2 for pea.

Plant measurements

Both maize and peas were planted at the same time and harvested 46 days after sowing (DAS), when maize plants were in mid-vegetative stage and pea plants were near the end of their vegetative stage. One plant at each end of each pot was discarded to remove edge effects. Plant height, leaf area and total shoot dry matter yield measurements were taken for both crops.

Derived measurements

RYT (de Wit & van den Bergh 1965) was used as a measure of resource complementarity and was expressed as RYT = (Yab/Yaa) + (Yba/Ybb), where Yab is yield per unit area of crop ‘a’ in mixture, Yaa is yield of sole crop ‘a’, Yba is yield of crop ‘b’ in mixture, and Ybb is yield of sole crop ‘b’. Yab/Yaa is relative yield (RY) of crop ‘a’ in mixture, and Yba/Ybb is relative yield of crop ‘b’ in mixture.

The competitive ability of a crop, relative to the other, was measured using a competitive ratio introduced by Willey & Rao (1980): CRa = (Yab/Yaa)/(Yba/Ybb) × Zba/Zab, where CRa is the competitive ratio of crop ‘a’ mixed with ‘b’, Zab is the proportion (number of plants) of crop ‘a’ in mixture with ‘b’, and Zba is the proportion of crop ‘b’ in mixture with ‘a’.

Data analysis

Analysis of variance ( anova) was carried out after testing for normal distribution, i.e. skewness and kurtosis (Snedecor & Cochran 1980). This was used to validate the use of anova and also to assess if data transformation was necessary or not. anova on plant height, leaf area and dry weight was performed as 24 factorial with four replications (i.e. 2 water stress × 2 pea cultivars × 2 root competition × 2 shoot competition; see Table 1). For relative yield total and competitive ratio, anova was performed as 2 × 2 × 3 factorial with four replications (i.e. 2 water stress × 2 pea cultivars × 3 types of competition – full competition, root competition only and shoot competition only; see Table 2). There was no first-order or second-order interaction between treatments, unless and otherwise stated in the results section.

Table 1. Results of anovas (mean squares and probability values) for dry weights (g plant−1), plant height (cm) and log leaf area per plant (cm2) of maize and pea. Bold numbers indicate significant differences between treatments
  Dry matter yield (g plant−1)Plant height (cm)Log leaf area per plant (cm2)
  MaizePeaMaizePeaMaizePea
 d.f.Mean square P Mean square P Mean square P Mean square P Mean square P Mean square P
  1. Values of skewness (Snedecor & Cochran 1980) at 0·05 level of probability are 0·459 for n = 70 and 0·492 for n = 60. Values for kurtosis (Snedecor & Cochran 1980) at 0·05 upper level of probability are = 3·87 for n = 75 and 3·99 for n = 50.

Replicates 3  83·99 0·0001  0·81 0·0001   19·730·6551  124·77 0·0092 0·25 0·0339 14·86 0·0001
Pea cultivar (PCV) 1  6·040·1299  2·69 0·0001   70·350·1708 2902·52 0·0001 0·70 0·0048  8·27 0·0001
Shoot competition (SC) 1  2·930·2887 0·200·0623  16·710·5011  35·400·27400·030·5307 0·080·4349
PCV × SC 1  3·910·2214 0·000·8226  21·040·4505  52·560·18390·190·1255 0·0020·8942
Root competition (RC) 1 135·58 0·0001  5·49 0·0001 1956·96 0·0001 3739·32 0·0001 4·82 0·0001  1·50 0·0013
PCV × RC 1  2·110·3666 0·100·1760   9·840·6052  25·500·35230·010·7812 0·00020·9622
SC × RC 1  0·000080·9956 0·310·0219   3·660·7524   0·390·90790·200·1241 0·070·4746
PCV × SC × RC 1  0·380·7008 0·100·1695  31·500·3566 147·010·02890·010·7716 0·180·2452
Water stress (WS) 1 372·63 0·0001  1·86 0·0001 7781·45 0·0001 6134·80 0·0001 8·07 0·0001  2·95 0·0001
PCV × WS 1  1·200·4961 0·110·1601   0·150·9490  73·530·11750·030·5708 0·230·1829
SC × WS 1  1·350·4700 0·040·4187   0·320·9260  17·220·44390·00020·9579 0·140·3007
PCV × SC × WS 1  1·600·4311 0·020·5866   0·540·9031   6·500·63730·080·3181 0·190·2299
RC × WS 1  0·470·6700 0·100·1862  26·650·3962  44·890·2188 1·14 0·0005  0·200·2137
PCV × RC × WS 1  24·96 0·0030  0·050·3606 111·040·0872 114·490·05250·310·0562 0·160·1008
SC × RC × WS 1  0·460·6734 0·100·1727  22·440·4359   0·050·96680·050·4289 0·080·4259
PCV × SC × RC × WS 1  1·040·5259 0·000·8980  22·680·4335   6·120·64730·040·4593 0·00040·9564
Error45  2·54  0·05   36·32   28·86 0·08  0·13 
R2   0·88  0·85    0·86    0·91 0·82  0·91 
CV  21·44 21·78    8·45   10·43 4·16  6·71 
Skewness (n = 64)   0·18  0·47    0·005    0·23 0·06 −0·21 
Kurtosis (n = 64)   0·04  3·42    0·48    0·31 2·08 −0·73 
Table 2. Results of anovas (mean squares and probability values) for relative yield total (RYT) and competitive ratio index of pea (CRp) per plant. Bold numbers indicate significant differences between treatments
  RYTCRp
 d.f.Mean square P Mean square P
  1. Values of skewness (Snedecor & Cochran 1980) at 0·05 level of probability are 0·532 for n = 50 and 0·558 for n = 45. Value for kurtosis (Snedecor & Cochran 1980) at 0·05 upper level of probability is 3·99 for n = 50.

Replicates 3  0·51 0·0005  0·160·4845
Pea cultivar (PCV) 1 0·0010·9209  1·36 0·0116
Competition (C) 2  1·99 0·0001  0·180·3896
PCV × C 2 0·090·2900 0·010·9447
Water stress (WS) 1  0·60 0·0052  0·720·0612
PCV × WS 1 0·030·5024 0·0011·0000
C × WS 2 0·140·1406 0·110·5782
PCV × C × WS 2 0·210·0550 0·050·7517
Error33 0·070·19  
R 2   0·76  0·34 
CV 19·2 46·3 
Skewness (n = 48)  0·38  0·83 
Kurtosis (n = 48)  0·45  1·10 

Results

Shoot dry matter yield

Maize

Neither shoot competition nor pea cultivar significantly affected the shoot dry matter yield of maize in the mixture, but there were significant effects (P < 0·001; Table 1) of both root competition and water stress. Root competition reduced biomass dry matter yield of maize by 33% (from 8·9 to 6·0 g plant−1) and water stress by 48% (from 9·8 to 5·0 g plant−1). There was also a significant (P < 0·01; Table 1) pea cultivar × root competition × water stress interaction (Table 3); in the absence of water stress, shoot dry matter yield of maize intercropped with pea cv. Bohatyr (compared with that intercropped with cv. Grafila) was reduced significantly (P < 0·01; Table 1) in the presence of root competition, but increased slightly in the absence of root competition. However, in the presence of water stress, shoot dry matter yield decreased significantly in the absence of root competition but was unaffected in the presence of root competition.

Table 3. Maize shoot dry matter yield (g plant−1) as affected by the interaction of water stress, root competition and pea cultivar. Each value is the mean of two shoot competition treatments
 − Water stress+ Water stress
 GrafilaBohatyrDifferenceGrafilaBohatyrDifference
  • **

    Significant at 1% level; NS = not significant. SED (0·05) to compare any individual values = 0·80.

− Root competition10·5811·85−1·27NS7·455·671·78**
+ Root competition 9·45 7·50 1·95**3·483·480·00NS
Difference 1·13NS 4·53** 3·97**2·19** 

Pea

Pea cv. Bohatyr produced 48% more (P < 0·001; Table 1) shoot dry matter (0·86 g plant−1) than cv. Grafila (0·12 g plant−1). Water stress reduced the shoot dry matter yield of peas by 27% (P < 0·001; Table 1), from 1·24 to 0·90 g plant−1, while root competition from maize reduced it by 43% (P < 0·001; Table 1), from 1·36 to 0·78 g plant−1. Pea shoot dry matter yield was not significantly affected by shoot competition, but there was a significant (P < 0·05; Table 1) shoot competition × root competition interaction (Fig. 2); the joint effect of root and shoot competition together was significantly (P < 0·05; Table 1) less than the sum of their individual effects.

Figure 2.

Pea shoot dry matter (DM) yield (g plant−1) as affected by the interaction of shoot and root competition: circle = no competition; diamond = root competition only; triangle = shoot competition only; and square = full competition. Each value is the mean of two pea cultivars and two water stress treatments. SED (0·05) to compare any two values is 0·09.

Plant height

Maize

Water stress reduced (P < 0·01; Table 1) the plant height of maize by 27% (from 82·4 to 60·3 cm) and root competition reduced it by 14% (from 76·9 to 65·8 cm). The effect of pea cultivar or shoot competition on plant height of maize was not significant.

Pea

Water stress reduced the plant height of pea (P < 0·001; Table 1) by 32%; in the presence of water stress pea height was 41·7 cm and in the absence of water stress it was 61·3 cm. The main effect of root competition and pea cultivar on pea height was significant (P < 0·001; Table 1). The main effect of shoot competition on pea height was not significant but there was a significant (P < 0·05; Table 1) shoot competition × root competition × pea cultivar interaction (Table 4). Shoot competition did not have a significant effect on the height of pea cv. Grafila in the presence or absence of root competition, whereas shoot competition reduced (P < 0·05; Table 1) the height of pea cv. Bohatyr in the presence but not in the absence of root competition. Root competition, however, decreased pea height of both pea cultivars in the presence or absence of shoot competition. Cultivar Bohatyr was significantly (P < 0·05; Table 1) taller than cv. Grafila regardless of the type of competition, on average 30% taller than cv. Grafila (58·2 cm compared with 44·7 cm).

Table 4. Pea height (cm) as affected by the interaction of pea cultivar, shoot competition (SC) and root competition (RC). Each value is the mean of two water stress treatments
 Cv. GrafilaCv. Bohatyr
 − SC+ SCDifference− SC+ SCDifference
  • *

    Significant at 5% level; NS = not significant. SED (0·05) to compare any individual values = 2·69.

− RC54·451·6 2·8NS65·465·00·4NS
+ RC34·738·2−3·5NS54·348·16·2*
Difference19·7*13·4*11·1*16·9*  

Leaf area

Maize

Leafy pea cv. Bohatyr decreased (P < 0·01; Table 1) the leaf area of maize by 21% compared with semi-leafless pea cv. Grafila. The leaf area of maize was similar in the presence or absence of shoot competition. The main effects of water stress (P < 0·001; Table 1) and root competition (P < 0·001; Table 1), and also the interaction between them (P < 0·001; Table 1) for maize leaf area, was significant (Table 5). Water stress reduced the leaf area of maize by 30% in the absence of root competition, but by 62% in the presence of root competition. Root competition reduced the leaf area of maize by 22% in the absence of water stress, but by 58% in the presence of water stress.

Table 5. Maize leaf area (cm2 plant−1) as affected by the interaction of root competition and water stress. Each value is mean of two pea cultivar and two shoot competition treatments. Values in parentheses are log transformed
 − Water stress+ Water stressDifference
  • *

    Significant at 5% level. SED (0·05) to compare any individual (log) values = 0·05.

− Root competition1471·4 (7·3)1029·2 (6·8)442·2 (0·5)*
+ Root competition1136·9 (7·0) 430·1 (6·0)706·8 (1·0)*
Difference 334·5 (0·3)* 599·1 (0·8)* 

Pea

Pea cv. Bohatyr produce 2·3 times more (P < 0·001; Table 1) leaf area (519·2 cm2) than cv. Grafila (230·3 cm2). The leaf area of pea declined (P < 0·001; Table 1) by 40% due to water stress, from 467·0 to 282·5 cm2. The effect of either root or shoot competition was non-significant.

Relative biomass yield total (ryt )

Water stress reduced RYT by 16% (P < 0·01; Table 2), from 1·47 to 1·24. The RYT of the intercrops was significantly (P < 0·001; Table 2) greater than 1·0 when peas and maize were in shoot competition (1·76), but not when they were in root competition (1·17) or full competition (1·13). Pea cultivars did not have a significant effect on RYT of mixtures.

Competitive ability of pea, relative to maize

The competitive ratio of leafy pea (cv. Bohatyr), when in full competition with maize, was 1·1 ± 0·15, indicating that it was as competitive as maize. In contrast, the competitive ratio of semi-leafless pea (cv. Grafila) was 0·8 ± 0·15, indicating that it was less competitive than either maize or leafy peas. Peas tended to have a lower root competitive ability than shoot competitive ability, but the difference was not significant. The competitive ability of peas tended to be greater when under water stress (CRp = 1·07) than when not under water stress (CRp = 0·82), but the difference was not significant (P > 0·05; Table 2).

Discussion

Shoot dry matter yield

Intercropping with peas, with both shoot and root competition, reduced the shoot dry matter yield of maize by 41%, while intercropping reduced the shoot yield of pea by 47%. In addition to this reduction in shoot yield, both maize and pea showed other characteristic responses to competition with reduction in leaf area and plant height.

Although root competition decreased the shoot dry weights of both maize and pea by decreasing their plant height and leaf area, shoot competition had no significant effect on these attributes, indicating that the effects of root competition were greater than those of shoot competition. The lack of significant effects by shoot competition indicates that soil resources, i.e. mineral nutrients and water, were more limiting than light; several reviews of the literature (Wilson 1988; Snaydon 1996) have shown that this is a common finding in experimental studies.

The greater effect of root competition compared with shoot competition in this study is not unexpected as the plants were harvested after only 46 days, and several studies have shown that root competition starts as soon as the root systems of one crop intermingles with that of the other, which usually occurs long before leaf canopies are sufficiently developed to allow competition for light (Pavlychenko 1940; Aspinall 1960). Moreover, the relative growth rate of most crops is seldom decreased until the light intensity falls to 50–60% of daylight (Blackman & Wilson 1951), as the leaf area expansion needs are met at lower levels of total daily radiation (Milthrope 1959).

Ryt

The RYT values of mixtures were generally greater than 1·0, regardless of the type of competition, the degree of water stress, and the pea cultivar used, indicating partial resource complementarity between maize and pea in this study. Most studies of legume/non-legume intercrops have given RYT values of greater than 1·0 (Trenbath 1974; Ofori & Stern 1987), which have been attributed to their use of different nitrogen sources and possibly to other differences in their root and shoot characteristics.

In this study, RYT values of the intercrops, in the absence of root competition (i.e. when only shoot competition occurred), averaged 1·76. This result is best considered in the light of the fact that the effects of root competition were greater than the effects of shoot competition, i.e. that soil resources were more limiting than aerial resources (see above). As a result, additive mixtures of the shoots of the maize and pea will give rise to less competition between the species than additive mixtures of their root systems; as both species are less affected by shoot competition, RYT values for shoot mixtures will be substantially greater than 1·0. In contrast, the mean RYT value of maize/pea intercrops averaged only 1·17, when in root competition only, and 1·13 when in full competition, i.e. much less than when in shoot competition (see above). This result follows from the fact that root competition was more intense than shoot competition. The RYT values of 1·17 and 1·13, for root competition only and full competition, are within the ranges commonly found in field studies of mixtures of legumes with non-legumes (Martin & Snaydon 1982; Tofinga, Paolini & Snaydon 1993) and probably reflect partial use of different nitrogen sources.

In studies where the effect of root and shoot competition have been separated for mixtures of legumes and non-legumes, results have been conflicting. Martin & Snaydon (1982) found values of RYT greater than 1·1 when root competition or full competition occurred but, in contrast, RYT values were not greater than 1·0 when both root and shoot competition occurred in mixtures of wheat and beans (Bulson 1991) or cereals and pea (Tofinga, Paolini & Snaydon 1993).

Although maize dominated the top of the canopy due to its taller height, there was no clear evidence that the tall maize required greater light than the pea plant: shoot competition reduced the biomass yield of pea by 10%, and biomass yield of maize by 5%. The fact that RYT values were almost the same for root competition only and full competition (root and shoot competition together) shows that not until below-ground deficiencies (water and nutrients) have been removed would light become a limiting factor (Snaydon & Harris 1981), partially or wholly. Pavlychenko & Harrington (1934) also reported that competition among intermingling roots of plants takes place much earlier than before the foliage of the plants starts shading one another.

RYT values were reduced when the intercrops experienced water stress. This shows that the degree of complementarity for water resources was greater when water was not limiting. However, the reduction in RYT value was not below 1·0, indicating spatial differences (Trenbath 1976) in the use of water resources, probably due to different rooting depths between maize and pea or easy access to the available water within the pots. Similarly, Whittington & O’Brien (1968) reported that mixtures of grass species (rye grass, meadow fescue and a triploid hybrid) that exploited different soil horizons produced yield greater than their respective sole crops. Ellern, John & Sagar (1970) also demonstrated the importance of deeper layers of the soil in the distribution of root systems of Avena fatua and A. strigosa.

Competitive ability

Leafy pea cv. Bohatyr had greater leaf area, plant height and shoot dry matter attributes, and consequently greater root and shoot competitive abilities, than semi-leafless pea cv. Grafila. This shows that leafy pea not only had a greater leaf canopy but also a more extensive root system than the semi-leafless pea. The difference in root competitive ability between the two cultivars was the same as their difference in shoot competitive ability, probably indicating that the difference in shoot morphology was reflected in root morphology.

Tofinga, Paolini & Snaydon (1993) also found that leafy pea cultivar Bohatyr was more competitive than semi-leafless pea cv. Countess. The semi-leafless pea cv. Grafila was, however, less competitive than maize in the present study. In contrast, Tofinga, Paolini & Snaydon (1993) found that even the semi-leafless pea cv. Countess had greater root and shoot competitive abilities than wheat or barley.

The competitive ratio of pea (CRp = 1·05) was only slightly greater than that of maize when both root and shoot competition were acting simultaneously, and slightly less than that of maize when in root competition (CRp = 0·84) or shoot competition only (CRp = 0·93). This shows that the competitive ability of pea depends on the combined effects of root and shoot competition.

The effect of root competition only on the competitive ability of pea relative to maize (as measured using the competitive ratio) was slightly but non-significantly greater than that of shoot competition only. In a review study by Wilson (1988), root competition had a greater effect than shoot competition, as measured using the competitive balance index, in 33 out of the 47 cases, whereas, shoot competition had a greater effect only in 14 cases.

Competition for water occurred between component crops in the mixture. The competitive ability of pea, relative to maize, was greater under water stress but less under no-water stress conditions. This supports the results of field experiments (Semere 1998) where the competitive ability of pea relative to maize was greater in the drier growing season of 1995, but less in the wet growing season of 1996, in the south-east of the UK. The decline in the competitiveness of maize subjected to water stress indicates the decrease in the ability of maize to make rapid use of its immediate water supply. Budelsky & Galatowitsch (2000) also found that control of water level fluctuation regimes (falling, rising or static) and competition were very important for the successful establishment and growth of wetland sedges (Carex lacustris) in reflooded wetland basins, especially during the first growing season.

Root competition from peas together with water stress reduced plant height, leaf area and dry matter yield of maize, and therefore might be expected to reduce further its ability to compete for light. However, there was no interaction between the effects of root competition and shoot competition on the competitive ability of the component crops, i.e. the effects of root and shoot competition were additive. Similar additive effects of root and shoot competition were found by Martin & Field (1984). However, the results of this experiment contradicts the hypothesis of Clements, Weaver & Hanson (1929) and the conclusion of Donald (1958) that root and shoot competition interact positively. King (1971) found no interaction, and Martin & Field (1987) found a negative interaction between root and shoot competition.

Because the experimental technique in the glasshouse does not reflect field conditions, caution is needed in extrapolating the results to a field situation. Nonetheless, in the glasshouse experiment of this study and in field experiments carried out by the same authors (Semere & Froud-Williams 1997; Semere 1998), leafy pea cv. Bohatyr was consistently more competitive than maize under water stress conditions. Semi-leafless pea cv. Grafila was less competitive than maize regardless of water stress treatments. Therefore, the environment created within the pots did not reverse the competitive relationship between the crops, although it intensified interspecific competition.

Acknowledgements

Part of this study was funded by the University of Reading and various educational trusts: the Society for the Protection of Science and Learning, the Airey Neave Trust in collaboration with the Laura Ashley Foundation, the Sir Richard Stapley Educational Trust, the Sidney Perry Foundation, the National Friendship Fund, the Leche Trust and the Professional Classes Aid Council. We are most grateful to all of them.

Received 16 March 1999; revision received 1 July 2000

Ancillary