Topology of a maize field

Abstract

Correlations were established between plant height and Cartesian position in a field of diverse maize (Zea mays) germplasm. The influence of the shade avoidance syndrome (SAS), a series of responses to lower photosynthetically active radiation (PAR) and red to far-red light ratio (R:FR) at high planting density, was detected by a steep increase of plant height from the edge to interior rows of the field. In addition, a gradual increase in height was observed across the field from east to west. We attribute this result to a R:FR gradient caused by sunlight laterally penetrating the stand at dusk. Furthermore, we hypothesize that the increased height of west-positioned plants may be analogous to responses induced by end-of-day FR (EOD-FR) treatments used by photobiologists to induce SAS in controlled environments. While preliminary, these results nevertheless suggest that a plant's position in a field will influence the impact of daily fluctuations in PAR and R:FR in modulating plant height and, potentially, other agronomically relevant traits.

To investigate the physiological and genetic components of the SAS,¹ several experimental designs have been developed to mimic the effects of plant crowding in the field.^2–5 In addition, experiments have been conducted in growth chambers using either FR reflected from low lying vegetation,⁶ low R:FR,⁷ or EOD-FR.^8,9 Examples of the application of EOD-FR treatments include the alteration of cold acclimation in red dogwood (Cornus stolonifera),¹⁰ maize seedling elongation,^11,12 and numerous effects in Arabidopsis (Arabidopsis thaliana).^13–16 The biological relevance of the EOD-FR to study the SAS remains controversial among photobiologists and to date, no naturally occurring EOD-FR response has been described for plants.

At twilight, the two periods of the day when the sun elevation is between −10° and +10°, refraction of sunlight by the atmosphere increases the relative amount of longer wavelengths that reach the earth's surface. The R:FR is reduced from approximately 1.19 during midday to 0.96 at dusk.¹⁷ The length of the two twilight periods and the extent of the R:FR reduction increase at higher latitudes, and also fluctuate seasonally, increasing in duration as daylight gets shorter in the fall.¹⁸ Reductions in the R:FR at both the beginning and end of the photoperiod, in addition to a lengthening of the photoperiod, were found to advance the seasonal developmental process of bud burst of silver birch (Betula pendula) plantlets when grown in controlled environments.¹⁹ In Arabidopsis, both low temperature and low R:FR at dusk can independently increase the expression of transcriptional activators leading to enhanced freezing tolerance.²⁰ Theses results suggest that prolonged low R:FR at the end of the growth season acts as a cue signaling the impending cold temperatures of winter.

To identify a potential naturally occurring effect of EOD-FR on plant development, we utilized field data collected for a quantitative trait loci analysis of plant height in maize (Edward Buckler, personal communication). We have previously measured seedling elongation responses to EOD-FR treatment in a growth chamber.¹¹ Thus, we examined here variation in plant height across a maize field in relation to the position and spectral quality of the sun.

Regression Analysis of Plant Height According to their Cartesian Field Coordinates

The effect of the X (W-E) and Y (S-N) Cartesian coordinates and their interaction (X * Y) was tested using least squares regression analysis and the following model: y = β₀ + β₁X + β₂Y + β₃XY + ε. The W-E axis (X) was divided into three sections; the outer four rows of the W and E borders (to control for field edges effect), and the remainder of the field. As shown in Figure 1A, both W and E borders had inverse slopes: +4.76 (p < 0.0001) and −4.12 (p = 0.0002) respectively. The two slopes are equivalent despite their heterogeneous composition (t = 0.44, p = 0.66). For the W border, both the X and Y components of the model were significant (t = 4.69 and 4.73 respectively, both p < 0.0001), but not their interaction (t = −1.88, p = 0.06). For the E border, only the X-axis component was significant (t = −3.82, p = 0.0002). The highly heterogeneous composition of the genetic material used, in combination with the relatively small sample size can possibly explain the significant Y-axis component of the W border. For the W border, the R² value for the model was 0.118 while the R² for the X and Y regressions were 0.054 and 0.055 respectively. At the E border, the R² of the model was 0.044 while the R² values of the X and Y regressions were 0.040 and 0.002 respectively. All R² values for the X and Y regressions were calculated by removing from the model the other Cardinal axis and the interaction.

Figure 1.

Regression plots of plant height measurements relative to their field Cartesian coordinates. Plant height is defined as the distance from the ground to the ligule of the flag leaf. Lines measured included both the NAM²¹ and IBM²² recombinant inbred populations, a panel of maize inbreds²³ and the B73 inbred line. Acting as check, B73 was randomly scattered throughout the field. (A) Plant height data plotted relative to their west-east (W-E) field coordinates (X-axis). The four outmost rows present at the W and E borders were analyzed separately to exclude the edge effect caused by greater PAR and R:FR. (B) Plant height data plotted relative to their south-north (S-N) field coordinates (Y-axis). Since each row is planted in a S-N orientation, any edge effect present at the S or N extremity of the field would be evident within the outmost S and N half-ranges of rows (first and last Y coordinates) and thus could not be detected with the available data. Regression analysis of B73 plant height relative to the (C) W-E coordinates and (D) S-N coordinates. Each row (oriented S-N) had a length of 2.7 m and planted at a density of 12 kernels. Rows were joined together by group of two, forming a range, followed by an empty space of 5.2 m. The distance between adjacent rows (X coordinates) of each half-range (Y coordinates) was 0.76 m. The number of observations recorded (n) is shown below each regression plot. A single plant corresponding to the median of each row was measured. The field had a uniform topography and unobstructed sunlight (42°N 43′ 34.748″, 76°W 39′ 14.065″).

We speculate that in the evening twilight, exposure to low R:FR laterally impinging from the W side of the field results in a plant height gradient along the W-E axis in the inner section of the field, with the tallest plants (or most SAS responsive) located at the W end of the inner section of the field. The analysis of the inner section of the field revealed that only the X component of the model was significant (slope of −0.28, t = −12.7, p < 0.0001) (Fig. 1A and B). The R2 value for the model is 0.034, while the R² values for the X and Y regressions are 0.034 and 0.0001 respectively. No S-N effect was detected, which suggests that the sun path inclination toward the south has no significant effect on plant height at this latitude (upstate New York, 42°N). This gradient also affects the plant height of the W and E borders. While shown to have equivalent slopes, the W border (as a group) is significantly taller than the E border (t = −8.02, p < 0.0001).

To remove variation resulting from genotypic differences, we performed a separate analysis of the B73 checks scattered across the field. As shown in Figures 1C and D, a significant effect was detected for the X-axis (t = −3.36, p = 0.001) but not the Y-axis (t = −0.24, p = 0.81). Furthermore, the X-axis regression slope of −0.24 is not significantly different to the X-axis slope for the whole population (t = 1.70, p = 0.089). The R² value for the model was 0.066, while the R² values for the X and Y regressions were 0.062 and 0.0012 respectively. The use of a single genetic background not only confirmed the W-E plant height gradient but also improved the proportion of the variation explained (R²) from 0.034 (whole population) to 0.062 (B73 only).

Conclusions

Using an extensive series of plant height measurements collected across a single maize field, we derived a series of regressions based on Cartesian field coordinates. Not only could a stiff increase in height, likely caused by the SAS, be detected at the W and E borders, but we also observed a W-E height gradient across the field. We hypothesize that this gradient is due to lower R:FR laterally penetrating the field from the W during the evening twilight period. However, we acknowledge that this analysis is based on unreplicated data and its interpretation speculative. Further experiments must be performed to confirm the presence and the extent of a physiological response to low R:FR at dusk. Additional characterizations of this response could include using phyB mutants as checks, which display a constitutive SAS response,^11,24 and thus would be predicted to not show a W-E gradient in plant height. It will also be important to consider field topography, fertilization, soil type and drainage in subsequent studies.

Nevertheless, these results suggest a more thorough assessment of light-induced phenotypic variations in a crop stand is warranted. Namely, that light responses should be factored into the design of experimental plots and analysis of precision agriculture data. Our prediction is that other traits such as grain yield would be slightly lower toward the W section of a field.²⁵ As SAS responses have been linked to increased herbivore susceptibility,²⁶ indirect yield reduction associated with SAS could accentuate the W-E gradient. It is tempting to speculate that the W-E gradient effect described here is mediated by a similar mechanism as the EOD-FR response.

Perspectives to: Dubois PG, Olsefski GT, Flint-Garcia S, Setter TL, Hoekenga OA, Brutnell TP. Physiological and genetic characterization of end-of-day far-red light response in maize seedlings. Plant Physiol. 2010;154:173–186. doi: 10.1104/pp.110.159830.