Mass selection of walnut rootstocks using response surface methods to correct for environmental trends

Abstract

Commercially grown fruit trees in the U.S. are almost invariably compound genetic systems with the top or scion cultivar budded or grafted to root systems derived from genetically unique zygotic seedlings. Although variation in tree size, yield, and cropping efficiency (CE) exists among trees within individual orchards, there have been no attempts to partition this variation in individual tree performance between the heterogeneity of zygotic seedling genotypes utilized as rootstocks and site-specific environmental variables such as soil type, fertility, and soil moisture, etc. We compared Field Blocking (FB), Moving Average (MA), and State Space (SS) procedures to infer environmental trends. Deviations from the estimated environmental effects that compose such trends were used to develop mass selection alternatives. FB was seen as a discrete approximation to an environmental response surface that varied regularly along tree rows. Regular trends within blocks for FB will result in a confounding of genetic and environmental effects and will decrease the precision of estimates of genetic potentiality based on block mean deviations. Trends estimated using the MA and SS procedures were very similar and appeared to minimize the confounding of genetic potentiality with the environmental response surface. Deviations from environmental trends estimated using the MA and SS procedures demonstrate that genetic variation of J. hindsii seedling rootstokes accounted for about 20% of the total variance in tree yield. Evidence for genotypic variation in CE was less clear.