Addressing huge spatial heterogeneity induced by virus infections in lentil breeding trials.

BACKGROUND: Spatial heterogeneity can have serious effects on the precision of field experimentation in plant breeding. In the present study the capacity of the honeycomb design (HD) to sample huge spatial heterogeneity was appraised. For this purpose, four trials were conducted each comprising a lentil landrace being screened for response to viruses. RESULTS: Huge spatial heterogeneity was reflected by the abnormally high values for coefficient of variation (CV) of single-plant yields, ranging 123-162 %. At a given field area, increasing the number of simulated entries was followed by declined effectiveness of the method, on account of the larger circular block implying greater intra-block heterogeneity; a hyperbolic increasing pattern of the top to bottom entry mean gap (TBG) indicated that a number of more than 100 replicates (number of plants per entry) is the crucial threshold to avoid significant deterioration of the sampling degree. Nevertheless, the honeycomb model kept dealing with variation better than the randomized complete block (RCB) pattern, thanks to the circular shape and standardized type of block that ensure the less possible extra heterogeneity with expanding the area of the block. CONCLUSIONS: Owing to the even and systematic entry allocation, breeders do not need to be concerned with the extra spatial heterogeneity that might induce the extra surface needed to expand the size of the block when many entries are considered. Instead, they could improve accuracy of comparisons with increasing the number of replicates (circular blocks) despite the concomitant greater overall spatial heterogeneity.

Wheat landraces are better qualified as potential gene pools at ultraspaced rather than densely grown conditions.

The negative relationship between the yield potential of a genotype and its competitive ability may constitute an obstacle to recognize outstanding genotypes within heterogeneous populations. This issue was investigated by growing six heterogeneous wheat landraces along with a pure-line commercial cultivar under both dense and widely spaced conditions. The performance of two landraces showed a perfect match to the above relationship. Although they lagged behind the cultivar by 64 and 38% at the dense stand, the reverse was true with spaced plants where they succeeded in out-yielding the cultivar by 58 and 73%, respectively. It was concluded that dense stand might undervalue a landrace as potential gene pool in order to apply single-plant selection targeting pure-line cultivars, attributable to inability of plants representing high yielding genotypes to exhibit their capacity due to competitive disadvantage. On the other side, the yield expression of individuals is optimized when density is low enough to preclude interplant competition. Therefore, the latter condition appears ideal to identify the most promising landrace for breeding and subsequently recognize the individuals representing the most outstanding genotypes.