Abstract:

Differences in spatial arrangement of plant organs strongly affect the interception of light within the canopy. Modelling light interception by maize leaves usually assumes a random distribution of leaf azimuth. Three-dimensional (3D) digitizing technology allows spatial data on plant architecture to be collected non-destructively and the movement of plant organs to be monitored dynamically. In order to simulate the course of possible changes in the plant geometrical structure during maize growth more accurately, some results were analyzed with circular data statistics using data from 3D digitizing.
Field measurements were conducted at the China Agricultural University (39º50’ N, 116º25’ E). Maize (Zea mays L. cv. ND108) was planted in north-south rows on June 29, 2003. There were two plant spacing treatments, both with rows spaced 0.6 m apart. Plant spacing of 0.3 m was used in the high plant density treatment and 0.6 m in the low plant density treatment. From the five-leaf stage to silking, measurements were carried out every 1–2 d since July 13, 2003 in the first experiment. Sixteen plants were selected for the study of the time course of leaf azimuthal movement and possible adaptive re-positioning of the successive leaves during and after their growth. At grain filling stage (75 d after sowing date), the second experiment (36 plants per plot) was conducted to assess the plant spacing effects on the distribution of the plant azimuthal planes and the deviation of leaves from the plane along the row.
An electromagnetic digitizer (3Space Fastrak Long Ranger, Polhemus) was used to measure the 3D coordinates of leaf midrib in situ for the taken plants in the field with two experiments. Data were collected when the plants were not affected by wind. Measurement points were selected to delineate the curve of midrib for each maize leaf. Then the azimuth of each leaf with respect to that of the row was computed according to collected 3D coordinates.
Azimuths of maize leaf were statistically analyzed with von Mises distribution which has two parameters (μ, k). The equation is , and in which . The parameter μ defines the reference direction of the distribution and k defines the deviation from that direction which k =0 represents the uniformity and the increasing k indicates the increasing concentration near that reference direction. Fisher test were used to compare the concentration parameters of the two circular distributions. Rayleigh test was used to test the uniformity of azimuthal plane of each individual plant amongst two treatments. The statistical results are as follows:
For any individual maize plant, there was an azimuthal shift from the bottom to the top. The parameter k tended to decrease with the rising of leaf position. The maximum azimuthal angle changed for a given leaf was about 150°. The final azimuth was fixed for 83% of the plant leaves about 3 d before the leaf fully expanded.
Plant azimuthal plane is essential to the study of the probability distribution of plant azimuth in field. The distributions of plant azimuthal plane for the two treatments were uniform. The plant spacing had an effect on dispersions of maize leaves from the plant azimuthal plane. From graphical inspection, a general trend of increasing dispersion with the rising of leaf position took place in both plant spacing treatments. For leaves ranked from 9 to 13, the distribution tended to be normal, with a rather low dispersion. For leaves 14 to 17, the dispersion was increased. For leaves 18 to 21, the angular distribution tended to be uniform. But for a single leaf, the concentration parameters were always higher in low density treatment than in high density treatment. The difference between treatments was significant at the 5% probability level for the leaves 9 to 13. This study provided a method of quantitative analysis and a description of the architecture in maize.

Key words: Maize, Circular data statistics, Canopy architecture, Azimuth, 3D digitizing