Crop Science 40:655-658 (2000)
© 2000 Crop Science Society of America
CROP BREEDING, GENETICS & CYTOLOGY
Plot Direction and Spacing Effects on Interplot Interference in Spring Wheat Cultivar Trials
F.R. Clarkea,
R.J. Bakera and
R.M. DePauwb
a Department of Plant Sciences, University of Saskatchewan, Saskatoon, SK S7N 5A8 Canada
b Semiarid Prairie Agricultural Research Centre, Swift Current, SK S9H 3X2 Canada
bob.baker{at}usask.ca
 |
ABSTRACT
|
|---|
Interplot interference can distort treatment estimates when genotypes differ for height. Two field arrangements were examined to determine if interplot interference could be reduced. One arrangement compared north–south vs. east–west row direction at Saskatoon in 1995 and 1996. The other experiment investigated the effects of separating plots with a row of spring-planted winter wheat (Triticum aestivum L.) at Regina and Swift Current in 1995 and 1996. Interplot interference was evaluated with two spring wheat cultivars differing for height, Oslo (short) and Glenlea (tall). Interplot interference caused a 12% yield reduction in Oslo in the north–south rows, which was significantly greater than the 7% yield reduction in the east–west rows. The 7% yield reduction when spring-planted winter wheat separated the plots was significantly less than the 18% yield reduction when plots were adjacent. This study was conducted at fairly high latitudes and the conclusions should be restricted to higher latitudes. We conclude that spring wheat field trials with plots differing for height may have less interplot interference if rows are oriented east–west and separated with winter wheat.
Abbreviations: K, Katepwa border • G, Glenlea • O, Oslo
|
INTRODUCTION
|
|---|
CLARKE ET AL. (1998) demonstrated that interplot interference can be a source of yield distortion in field trials containing genotypes differing for height in the short growing season on the Canadian prairies. The interrow and interplot spacing was 0.3 m and plot rows were oriented north–south. Spring-planted winter wheat has been used to separate plots and reduce interplot competition. May and Morrison (1986) concluded that as long as the separation method is not more competitive, the separation method should not alter yield selection. Barley (Hordeum vulgare L.) and spring wheat plots separated with spring-planted winter wheat were less competitive than when flanked by the same genotype or when a barley plot was flanked by wheat or a wheat plot flanked by barley (May and Morrison, 1986). However, increased space between plots may also increase heterogeneity within blocks (Spitters, 1979; Federer and Basford, 1991).
When row direction was indicated in studies that reported interplot interference in field trials, the row direction was generally north–south (Austin et al., 1977; Austin and Blackwell, 1980; Kempton and Lockwood, 1984; Kempton et al., 1986; Clarke et al., 1998). However, Kiesselbach (1919) and Jensen and Federer (1964) reported interplot interference in trials with east–west rows as well with north–south rows, and Fisher (1979) and Kempton et al. (1986) reported interplot interference when rows were east–west. Baker and Rossnagel (1988) reported significant interplot interference in three of four north–south tests, two with wheat and one with barley, and not in the four east–west tests. Baker and Meyer (1966) demonstrated that during the morning and late afternoon, north–south rows admitted more light than east–west rows. At Cambridge, Kempton et al. (1986) calculated that plots with north–south rows would receive a 0.7% net loss of radiation for each centimeter shorter than its flanking plots, and that the reverse would occur for each centimeter taller than its flanking plots. However, with east–west rows, the net loss (-0.5% cm-1) in radiation for shorter plots was less than the net gain (0.6% cm-1) for taller plots.
We examined whether interplot interference is a concern when plots that differ in height are separated by spring-planted winter wheat vs. those that are not and whether interplot interference is greater when rows are north–south than when they are east–west.
|
Materials and methods
|
|---|
In 1995 and 1996, the row-direction experiment was conducted at the University of Saskatchewan Kernen Farm at Saskatoon, and the plot-separation experiment was conducted at the Semiarid Prairie Agricultural Research Centre at Swift Current and Regina. Oslo and Glenlea, two hexaploid wheat cultivars that represented the extremes in height grown in Western Canada, were used for the two interplot interference treatments. Glenlea (Evans et al., 1972) is taller than Oslo (Graf et al., 1990). Both treatments consisted of three four-row plots, with the pure-stand treatment a four-row plot of Oslo (O) flanked by Oslo, O-O-O, and the other interference treatment Oslo flanked by Glenlea (G), G-O-G. The difference in yield of Oslo in the center plot of G-O-G and of Oslo in the center plot of O-O-O is the interference effect. The seed was planted into moist soil to a depth of 
2 cm. Height was measured, the entire plot was harvested, and the grain was dried and weighed to determine yield.
At Saskatoon, interplot and interrow spacing was 0.3 m, the plots were 3.7 m long and the plots had 1.8-m cultivated alleys between ranges. The plots were seeded at 270 seeds m-2. Fertilizer (11-51-0 N-P-K) was applied at a rate of 56 kg ha-1 with the seed to provide 6.2 kg ha-1 N and 12.4 kg ha-1 P. A herbicide that contained bromoxynil (3,5-dibromo-4-hydroxyphenyl cyanide) and MCPA ([(4-chloro-o-tolyl)oxy]-acetic acid) was applied at the four- to five-leaf stage for weed control. At Swift Current and Regina, the plots were 3 m long with four rows 0.23 m apart, seeded at 250 seed m-2, and trimmed to 3 m with 2-m spring-planted winter wheat alleys between ranges. Nitrogen (34-0-0) was broadcast at 45 kg ha-1, and N and P (11-51-0) were broadcast at 20 kg ha-1 prior to planting. The herbicide tralkoxydim (2-[1-(ethoxyimino)propyl]-3-hydroxy-5-(2,4,6-trimethylphenyl)-2-cyclohexen-1-one) was applied at the four- to five-leaf stage followed at least 4 d later with the herbicide bromoxynil and MCPA.
In the row-direction experiment, two plot row directions, north–south rows and east–west rows, were in combination with the two interplot interference levels (Fig. 1)
. Each of three blocks contained two plot row directions and the blocks were adjacent on their long side. Each year, the first row direction in one block was randomly chosen and the alternate row direction became the next row direction in that block. The row directions in the remaining two blocks alternated so adjacent blocks never had the same row direction side by side. Each row direction within each block contained two replications of the two interplot interference levels arranged in a randomized complete block. The interference levels in the first replication were randomly chosen and the alternate order was applied to the second replication. Both replications in each row direction contained a single range of plots that was bordered on each end with a plot of `Katepwa' spring wheat, with plots adjacent along their long side.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 1 Field plan for the row-direction trial at Saskatoon in 1995 showing three blocks, each with north–south and east–west row directions, and each direction with two replications. An example of the treatments is shown, where K = Katepwa border, G = Glenlea, and O = Oslo
|
|
For the plot-separation experiment, each location contained two subexperiments, each with plot rows oriented north–south. In the subexperiment designed to reduce interplot interference, plots were separated by two rows of spring-planted winter wheat 0.05 m apart, and plots were 0.23 m on either side of the two winter wheat rows. Thus, adjacent plots were 0.23 + 0.05 + 0.23 = 0.51 m apart. In the other subexperiment, spacing between plots was the same as spacing between rows within plots (0.23 m). Each subexperiment was arranged as a randomized complete block with 12 replications of the two interplot interference treatments. In each subexperiment at each location in each year, the two interference treatments were randomized in the first replication, and the alternate order was used in the next replication. The above process was repeated for each remaining pair of replications. At most, only two adjacent replications could have the same treatment order.
Mixed-models analysis was performed with PROC MIXED in SAS (version 6.12; SAS Institute, 1992). The MODEL statement contained the fixed factors and the RANDOM statement contained the random factors. Satterthwaite's approximate degrees of freedom for the appropriate error term were calculated by including /DDFM=SATTERTH as an option in the MODEL statement. Weighting of heterogeneous errors from the different environments was included with the GROUP option in the REPEATED statement. Analysis of each experiment involved five main effects. For the row-direction experiment, these were years (Y), blocks (B) in years, directions (D), replications (R) in blocks and directions, and interference treatment (T). For the plot-separation experiment, these were years, locations (L), replications in years and locations, plot-separation method (S), and interference treatment.
For the row-direction experiment, we analyzed each block in each year separately with the SAS statements:
and calculated the mean effect of interference for the two replications with plots in the north–south direction, the mean effect of interference for the two replications with plots in the east–west direction, and the difference between these two estimates of interference. One-tailed t tests were performed to determine if the mean interference effects were greater than zero. We analyzed the combined data across years of the row-direction experiment with:
For the plot-separation experiment, we analyzed each location in each year separately with:
and calculated the mean effect of interference for the 12 replications with plots separated by winter wheat, the mean effect of interference for the 12 replications with plots not separated by winter wheat, and the difference between these two estimates of interference with their standard errors. We analyzed the combined data across years and locations of the plot-separation experiment with:
Estimates for narrow and broad inference spaces (McLean et al., 1991) were calculated for each combined analysis.
|
Results and discussion
|
|---|
Height of Oslo was not significantly affected by taller flanking plots of Glenlea in the row-direction and plot-separation experiments (data not shown), which agrees with results of Smith et al. (1970), Hamblin and Donald (1974), and Clarke et al. (1998). In the row-direction and plot-separation experiments, Glenlea averaged 41% taller than Oslo (Table 1)
.
In the row-direction experiment, analyses of individual blocks in each year revealed significant reduction of Oslo when flanked by plots containing Glenlea in seven of 12 comparisons (Table 2) . When rows were oriented north–south, the level of interference on yield of Oslo (average = -59 g m-2) was greater than when rows were oriented east–west (average = -35 g m-2). The difference between row directions was significant in one block in 1995 and one block in 1996 (Table 2). For the 2 yr of this experiment, with rows oriented north–south, the average yield reduction of 12% for Oslo (from a pure stand yield of 483 g m-2) was significant (-59 ± 7 g m-2, P < 0.01). However, when viewed in the context of average performance over a population of years from which these two are a sample (broad inference space of McLean et al., 1991), this reduction is judged not significant (Table 2). When rows were oriented east–west, the overall 7% yield reduction of Oslo was significant (-35 ± 7 g m-2, P < 0.01) in the narrow inference space of these 2 yr, but not significant in the broad inference space (-35 ± 20 g m-2). More variation in interplot interference occurred in 1995. On average, interference was almost two-thirds greater (-24 ± 10 g m-2, P < 0.05) when rows were oriented north–south than when oriented east–west.
View this table:
[in this window]
[in a new window]
|
Table 2 Mean response and standard errors of Oslo to interplot interference from Glenlea in the row-direction experiment at Saskatoon in 1995 and 1996. Interplot interference effects are averaged across the two replications in each block
|
|
In an experiment at Saskatoon with rows oriented north–south, the reduction from pure stand yield of Oslo was 17% in 1993 and 10% in 1994 (Clarke et al., 1998). The reduction in interplot interference supports the indications of a pattern of greater interference when rows are oriented north–south than east–west observed by Baker and Meyer (1966), Kempton et al. (1986), and Baker and Rossnagel (1988).
With the plot-separation experiment at Regina and Swift Current in 1995 and 1996, analysis of each subexperiment revealed that interplot interference tended to be significant when spring-planted winter wheat separated plots and when no spring-planted winter wheat separated plots in each environment, with the exception of Swift Current in 1996 (Table 3)
. The level of interplot interference was significantly greater when plots were not separated. Across years and locations, interplot interference significantly reduced yield by 18% (-72 ± 5 g m-2, P < 0.01) from pure stand yield (410 g) when plots were not separated with spring-planted winter wheat, but if applied to future years this reduction was not significant (-72 ± 18 g m-2). The overall 7% reduction in yield with separation was significant (-27 ± 5 g m-2, P < 0.01) if restricted to Regina and Swift Current in 1995 and 1996 because of minimal residual variation, but was not significant (-27 ± 18 g m-2) when applied to future years. The reduction in yield from interplot interference was significantly greater (-45 ± 7 g m-2, P < 0.01) when plots were not separated with winter wheat than when they were. Yield loss from interplot interference for the different methods of plot separation was inconsistent among the 2 yr and two locations (Table 3). The plot-separation experiment conducted at Swift Current and Regina in 1995 and 1996 indicated that interplot interference can occur when plots are separated with spring-planted winter wheat, but to a lesser degree than when adjacent. May and Morrison (1986) demonstrated that spring wheat or barley genotypes bordered with spring-planted winter wheat yielded 4 to 21% more than pure-stand yield. This yield increase could add a further distortion to yield comparisons if some genotypes responded differently.
View this table:
[in this window]
[in a new window]
|
Table 3 Mean response and standard errors of Oslo to interplot interference from Glenlea in the plot-separation experiment at Swift Current and Regina in 1995 and 1996. Interplot interference effects are averaged across the 12 replications in each subexperiment
|
|
We observed significant yield reduction due to interplot interference in each field experiment and in each environment. However, the level of variation in observed interference prevents a reliable prediction of average interference in other experiments or environments. The row-direction and plot-separation experiments indicated that interplot interference was greater in north–south rows than in east–west rows and was greater when plots were adjacent than when plots were separated with spring-planted winter wheat. If the reductions in interplot interference under these conditions are consistent, then plots with east–west rows and plot separation should show the lowest interplot interference. These conclusions should be restricted to higher latitudes similar to those where the research was conducted.
|
ACKNOWLEDGMENTS
|
|---|
We gratefully acknowledge the technical assistance of Bill Dougall and Ken Saretsky, and the help of numerous staff members of the Arid Prairie Wheat Program at the Agriculture and Agri-Food Canada Agriculture Semiarid Prairie Agricultural Research Centre at Swift Current. Part of the field work was financed through a Natural Sciences and Engineering Research Council of Canada Individual Research Grant for R.J. Baker and part by the Arid Prairie Wheat Program. The advice of Dr. J.M. Clarke is gratefully acknowledged.
|
NOTES
|
|---|
Part of a Ph.D. thesis submitted by F.R. Clarke.
Received for publication April 23, 1999.
|
REFERENCES
|
|---|
- Austin R.B., Blackwell R.D. Edge and neighbour effects in cereal yield trials. J. Agric. Sci. (Cambridge) 1980;94:731-734.
- Austin R.B., Ford M.A., Edrich J.A., Blackwell R.D. The nitrogen economy of winter wheat. J. Agric. Sci. (Cambridge) 1977;88:159-167.
- Baker D.N., Meyer R.E. Influence of stand geometry on light interception and net photosynthesis in cotton. Crop Sci. 1966;6:15-19.
- Baker R.J., Rossnagel B.G. Interplot competition between wheat or barley cultivars of differing heights. Can. J. Plant Sci. 1988;68:1129-1132.
- Clarke F.R., Baker R.J., DePauw R.M. Interplot interference distorts yield estimates in spring wheat. Crop Sci. 1998;38:62-66.[Abstract/Free Full Text]
- Evans L.E., Shebeski L.H., McGinnis R.C., Briggs K.G., Zuzens D. Glenlea red spring wheat. Can. J. Plant Sci. 1972;52:1081-1082.
- Federer W.T., Basford K.E. Competing effects designs and models for two-dimensional field arrangements. Biometrics 1991;47:1461-1472.
- Fisher R.A. Are your results confounded by intergenotypic competition?. In: Ramanujam S., ed. Proc. Int. Wheat Genetics Symp., 5th. New Delhi, India. 23–28 Feb. 1978. New Delhi, India: Indian Soc. Genet. Plant Breeding, Indian Agric. Res. Inst, 1979:767-777.
- Graf R.J., Hucl P., Smith J., Song L.S.P. Oslo red spring wheat. Can. J. Plant Sci. 1990;70:299-302.
- Hamblin J., Donald C.M. The relationships between plant form, competitive ability and grain yield in a barley cross. Euphytica 1974;23:535-542.
- Jensen N.F., Federer W.T. Adjacent row competition in wheat. Crop Sci. 1964;4:641-645.[Free Full Text]
- Kempton R.A., Lockwood G. Inter-plot competition in variety trials of field beans (Vicia faba L.). J. Agric. Sci. (Cambridge) 1984;103:293-302.
- Kempton R.A., Gregory R.S., Hughes W.G., Stoehr P.J. The effect of interplot competition on yield assessment in triticale trials. Euphytica 1986;35:257-265.
- Kiesselbach T.A. Experimental error in field trials. Agron. J. 1919;11:235-241.[Free Full Text]
- May K.W., Morrison R.J. Effect of different plot borders on grain yields in barley and wheat. Can. J. Plant Sci. 1986;66:45-51.
- McLean R.A., Sanders W.L., Stroup W.W. A unified approach to mixed linear models. Am. Stat. 1991;45:54-64.
- SAS Institute. 1992. SAS technical report P-229. SAS/STAT software: Changes and enhancements. Release 6.07. SAS Inst., Cary, NC.
- Smith O.D., Kleese R.A., Stuthman D.D. Competition among oat varieties grown in hill plots. Crop Sci. 1970;10:381-384.[Abstract/Free Full Text]
- Spitters C.J.T. Competition and its consequences for selection in barley breeding. Agricultural Res. Rep. 893. Wageningen, the Netherlands: Centre for Agricultural Publishing and Documentation, 1979.
TOP
ABSTRACT
INTRODUCTION
