The relationship between yield, mainstem number, and tuber number in five maincrop and two early-maturing cultivars

Using data from an 11-year period, total and marketable yield were related, using regression analysis, to mainstem and tuber number for five maincrop varieties (Russet Burbank, Shepody, Atlantic, Norchip, and Norland) and two early-maturing varieties (Conestoga and Carlton), at irrigated and dryland sites in western Canada. Mainstem number by itself was a poor predictor of total and marketable yield. Highly significant regressions of yield on the derived variable tuber number per mainstem were evident for all cultivars grown at the dryland sites. In general, the proportion of the total variation in yield accounted for by the regression models was much higher at the dryland than at the irrigated sites, indicating that the relationship between yield and its predictor variables was more complex for irrigated sites. While the study suggests that the use of tuber number per mainstem may be useful as an early-season predictor of tuber yield for cultivars, particularly under dryland conditions, a model using the predictor variables mainstem and tuber number will generally be superior due to the additional information it provides when competitive effects are present for light energy within the canopy or for assimilates among developing tubers. Contour plots provide a useful way to study the yield response to changes in tuber number at specific stem density levels.