DEMONSTRATION AND EVALUATION TRIALS (TABLE 3)

Analyses of data from seven sites

Unless otherwise specified, where significant differences were found from the ANOVA tests these were at the F = I % level. Where no significant differences were found these were at the F = 5% level.

i. Between treatments - the most prominent factor affecting vegetative growth, as determined by plant height, appears to be manure. The stools receiving five debes of cow manure at planting (Treatments FC2, FC and F) had plants which clearly were taller than those from stools which had received only one debe of manure. In the ANOVA, the stools with the low manure, no pesticide treatment (Treatment O) had plants which were significantly shorter than those on the high manure stools. The stools with low manure and pesticide at planting (Treatment C) also had plants which were significantly shorter (F = 5%) than those in two of the high manure treatments (FC2 and F). Following ANOVA, a Duncan's Multiple Range Test (DMRT) was applied to the mean heights in each of the experimental plots (omitting the results from Rubale, where there were no low manure plots). The 36 plots were ranked and each was allotted a score on the basis of 1 point for the first ranked plot to 36 for the lowest ranked. Cumulating the scores for each treatment gave: FC2,65 points; F, 69; FC, 83; C2, 134; C, 132; and O,181.

ii. Between sites (omitting Rubale) - the ANOVA showed significant separation of sites into Izimbya; Kangantebbe and Kalonge; Kalema and Kaigara; and lbaraizibu. As above, a DMRT enabled a ranking of all 36 plots and allotting of scores which, when cumulated, gave; Izimbya, 59 points; Kangantebbe, 79; Kalonge, 86; Kalema, 128; Kaigara, 143; and Ibaraizibu, 169.

iii. Within treatments - the ANOVAs generally reflected the overall ranking of sites (as in the between sites analysis), except that no significant differences were found between sites in the FC treatment and only one site (Kaigara) differed from the others in the C2 treatment.

Analyses of within site results

Ibarazibu - the low manure plots had the lower mean heights, with C significantly poorer than all except O, but C2 was only marginally poorer than the high manure plots.

Kalema - FC2 was significantly better than C and O; O was significantly poorer than all except C.

Kalonge - there were no significant differences between treatments, although the three high manure plots had the greater mean heights.

Izimbya - treatments FC2 and O were significantly different from each other but neither was separable from the other four treatments.

Rubale - there were only high manure treatments and these did not differ significantly.

Kangatebbe - treatment O was significantly poorer than all other treatments.

Kaigara - the three high manure treatments were similar and all were significantly better than the low manure treatments which in their turn were separable.

Analysis according to manuring levels

i. Within the high manuring treatments - a DMRT ranking of the 21 plots, from all seven sites, followed by cumulation of site scores gave; Izimbya, 15 points, mean height 334.0 cm; Kalonge, 18 points, 321.0 cm; Kangantebbe, 23 points, 314.2 cm; Rubale, 28 points, 309.9 cm; Kaigara 42 points, 295.4 cm; Kalema 44 points, 292.7 cm; and Ibaraizibu, 60 points, 273.9 cm. Izimbya and Ibaraizibu differed significantly from each other but neither differed significantly from the other five sites. Cumulation of ranking points on a treatment basis showed no significant differences between the treatments at any one site. The pesticide usage made no apparent difference to the growth of the plants on an overall or within site basis.

ii. Within the low manuring treatments - a DMRT ranking of the 18 plots, followed by cumulation of the site scores, gave; Izimbya, 11 points, mean height 296.3 cm; Kalonge, 22 points, 271.8 cm; Kangantebbe, 18 points, 270.2 cm; Kalema 31 points, 247.8 cm; Ibaraizibu, 44 points 217.4 cm; and, Kaigara 44 points, 206.9 cm. In this analysis, Izimbya differed significantly from Ibaraizibu and Kaigara, but the other three sites did not differ significantly from those three. Cumulation of ranking points on a treatment basis showed no significant difference between treatments. Overall, treatment O had the shortest plants. This seemed to reflect an effect of pesticide use at Ibaraizibu, Kalema and Kangantebbe; with treatment O being significantly the poorest at the latter two sites. At Kalonge, Izimbya and Kaigara no effect of pesticide use was shown.

NEXT

href="\bananas2000\analyse1.htm"