Previous PageTable Of ContentsNext Page

Quantifying loss in grain yield due to poor plant stand in sunflower

L.J. Wade, C.G. Qiao, A.C.L. Douglas and J.R. Agnew

Department of Primary Industries, PO Box 81, Emerald QLD 4720

Establishment of sorghum and sunflower was surveyed (1), using a multiple regression technique (2,3). The survey found loss in grain yield could be substantial. Because the technique had not been independently evaluated, experiments were conducted: (i) to compare estimates of grain yield from the regression with direct measurements from header harvesting, (ii) to add a term for seedling vigour, and (iii) to quantify the minimum gap or clump for significant loss in grain yield of sunflower.


Sunflower cv. Hysun 41 was planted in 1 m rows at the Emerald Research Station, on 12 September 1990. Time of emergence of individual plants was recorded from three adjacent rows of crop. By hand-thinning after emergence, 18 plant density by uniformity treatment combinations were obtained. The treatments, replicated three times, comprised three plant densities (3, 6 and 12 plants/m2) and six levels of clumping (1, 2, 4, 6, 8 and 12 plants/clump). At maturity, heads of individual plants were harvested and weighed, following measurement of inter-plant spacing’s. Grain weights were obtained for some heads, the remaining grain weights being estimated by regression. Grain weight per plant was thick regressed against plant spacing and plant spacing uniformity (2,3). Multiplying the equation by plant density provided a direct relationship with grain yield. Inter-plant spacing was recorded for the rows to be header harvested. Mean values of plant spacing and uniformity were calculated for these rows and substituted into the regression, to obtain a prediction of grain yield for each treatment. Predicted and observed grain yields were compared. Time to emergence and its interactions with plant spacing and uniformity were added to the regression. Grain yields from the 18 treatment combinations were compared.

Results and discussion

Agreement between predicted and observed grain yields was good (R2 = 0.92, n = 18), confirming that the technique was valuable for quantifying loss in grain yield due to a poor plant stand. Nevertheless, the regressions for individual treatments accounted for only about 20% of the variation. A similar accountability was obtained in the establishment survey (1). Since these regressions only considered the impact of plant spacing geometry, low accountability was not surprising (2,3). Accountability improved from 28 to 44%, when time to emergence and its interactions with plant spacing and uniformity were added. In variable plant stands, gaps greater than 2 m resulted in significant loss in grain yield. Work is now proceeding to develop a technique whereby growers may evaluate the quality of a plant stand, and decide whether replanting is warranted. This project is funded by the Oilseeds Research Committee.


Radford et al. 1989. QDPI Project Report Q089027.

Wade, L.J., Norris, C.P., and Walsh, P.A. 1988. Aust. J. Exp. Agric. 28, 617-622.

Wade, L.J. 1990. Aust. J. Exp. Agric. 30, 251-255.

Previous PageTop Of PageNext Page