8.7 Unreplicated early generation variety trial 104At present asreml cannot average over pol() terms so we must specify the value of Columnat which the predictions are to be formed; in this case we choose to form varietal predictionsat the average value of Column, that is, 5.5.> wheat4.pv$predictions$"Variety:Column":$"Variety:Column"$pvals:Column Variety predicted.value standard.error estStatus1 5.5 1 2917.178 179.28821 Estimable2 5.5 2 2957.741 178.76889 Estimable3 5.5 3 2872.762 176.98813 Estimable4 5.5 4 2986.473 178.74259 Estimable...522 5.5 522 2784.768 179.15427 Estimable523 5.5 523 2904.942 179.53841 Estimable524 5.5 524 2740.034 178.84664 Estimable525 5.5 525 2669.956 179.24457 Estimable526 5.5 526 2385.981 44.21617 Estimable527 5.5 527 2697.068 133.44066 Estimable528 5.5 528 2727.032 112.26513 Estimable529 5.5 529 2699.824 103.90633 Estimable530 5.5 530 3010.391 112.30814 Estimable531 5.5 531 3020.072 112.25543 Estimable532 5.5 532 3067.448 112.66467 Estimable$"Variety:Column"$avsed:[1] 245.8018Note that the replicated check lines have lower SEs than the unreplicated test lines. Therewill also be large differences in SEDs. Rather than obtaining the large table of all SEDs,the prediction could be done in parts if the interest was to to examine the matrix ofpairwise prediction errors of check varieties, for example.> wheat5.pv wheat6.pv wheat6.pv$predictions$"Variety:Column":$"Variety:Column"$pvals:Notes:- weed evaluated at average value of 0.459701- pol(Column, -1) evaluated at 5.500000- units terms are ignored unless specifically included- mv is averaged over fixed levels- pol(Column, -1) is included in the prediction- weed is included in the prediction- (Intercept) is included in the prediction- units is ignored in this prediction- mv is ignored in this predictionColumn Variety predicted.value standard.error est.status1 5.5 526 2385.981 44.21617 Estimable
2 5.5 527 2697.068 133.44066 Estimable3 5.5 528 2727.032 112.26513 Estimable4 5.5 529 2699.824 103.90633 Estimable5 5.5 530 3010.391 112.30814 Estimable6 5.5 531 3020.072 112.25543 Estimable7 5.5 532 3067.448 112.66467 Estimable$"Variety:Column"$sed:8.8 Paired Case-Control Study 105[,1] [,2] [,3] [,4] [,5] [,6] [,7][1,] NA 129.1900 107.2885 98.20983 107.4884 107.0353 107.4493[2,] 129.18995 NA 165.5508 159.10100 165.5290 165.4920 165.9755[3,] 107.28848 165.5508 NA 141.34763 148.5679 149.5352 148.2006[4,] 98.20983 159.1010 141.3476 NA 143.0536 142.1024 143.3973[5,] 107.48844 165.5290 148.5679 143.05365 NA 144.3514 149.5803[6,] 107.03532 165.4920 149.5352 142.10244 144.3514 NA 149.5490[7,] 107.44926 165.9755 148.2006 143.39734 149.5803 149.5490 NA$"Variety:Column"$avsed:min mean max98.20983 139.90452 165.975538.8 Paired Case-Control StudyThese data are from an experiment conducted to investigate the tolerance of rice varietiesto attack by the larvae of bloodworms. The data have been kindly provided by Dr. MarkStevens, Yanco Agricultural Institute. A full description of the experiment is given byStevens et al. [1999]. Bloodworms are a significant pest of rice in the Murray and Murrumbidgeeirrigation areas and damage can result in poor establishment and substantialyield loss.The experiment commenced with the transplanting of rice seedlings into trays. Each traycontained a total of 32 seedlings and the trays were paired so that a control tray (nobloodworms) and a treated tray (bloodworms added) were grown in a controlled environmentroom for the duration of the experiment. After this, rice plants were carefullyextracted, the root system washed and root area determined for the tray using an imageanalysis system described by Stevens et al. [1999]. Two pairs of trays, each pair correspondingto a different variety, were included in each run. A new batch of bloodwormlarvae was used for each run. A total of 44 varieties was investigated with three replicatesof each. Unfortunately the variety concurrence within runs was less than optimal.Eight varieties occurred with only one other variety, 22 with two other varieties and theremaining 14 with three different varieties.The following subsections present an exhaustive analysis of these data using equivalentunivariate and multivariate techniques. It is convenient to use two data frames, one foreach approach. The univariate data frame> rice names(rice)[1] "Pair" "rootwt" "Run" "sqrtroot" "Tmt" "Variety"has factors Pair, Run, Variety, Tmt and variates rootwt and sqrtroot. The factorPair labels pairs of trays (to which varieties are allocated) and Tmt is the two levelbloodworm treatment factor (control/treated).