The experience gained from the surveys led to the setting up of a series of simple "On-farm Demonstration and Evaluation trials", spread widely in Bukoba, Muleba and Karagwe Districts, and two Factorial trials in the area of the most degraded shambas in Bukoba District.

The initial aim was to establish trials on five sites within each of the four administrative Districts (Bukoba, Muleba, Karagwe and Ngara) of the Region in which bananas are the staple food crop. Shambas with bananas in poor condition were chosen for siting the trials. The preference was for a site measuring 27 x 27 m, which would permit the planting of 64 banana stools, with an inter-stool spacing of 3 m, in a square of eight rows each with eight stools. When such a shape was possible, the site was subdivided to give four plots each four stools square, i.e. 16 stools per plot. Each plot received a different treatment and there was no replication of treatments at a site. The existing bananas were dug up and the corm material chopped; this it was hoped would give a fairly even inoculum of nematodes, if present. Planting holes 60 cm deep and 90 cm wide were dug and the planting material in all holes was of the pared corm type; with one corm per hole as the parent of each stool. It was recognised that it might not be possible to obtain sufficient corms of a single clone at most of the sites but the element of clonal variety was felt to have sufficient scientific merit to offset this as a problem. The measure of cow manure was a common 20 l container known as a debe.

The four treatments at the time of planting were:

• Plot O - planting with no pesticide and with only one debe of cow manure per stool.
• Plot F - planting with no pesticide but with five debes of cow manure per stool.
• Plot C - planting with carbofuran (60 g Furadan 5G granules) at a dose of 3 g a.i. plus one debe of cow manure per stool.
• Plot FC - planting with carbofuran (60 Furadan 5G granules) at a dose of 3 g a.i. plus five debes of cow manure per stool.

The single debe in the plots O and C was felt desirable to promote the germination and initial growth of the corms. If a site was not level, the layout was so that plot O was at the highest corner and FC at the lowest. F was alongside O and uphill of FC. C was downhill from O and alongside FC.

Six months after planting the plots C and FC were subdivided and one half, eight stools, of each plot was given a further dose of carbofuran at 3 g a.i. per stool, broadcast around the stool. The subplots were designated C2 and FC2. The simplicity of the trial design and layout was decided as being the most manageable in a situation where routine site management would have to be the responsibility of a village level extension worker, with only a basic agricultural training. The original intention was for data on germination, plant vigour, survival, flowering and yield to be gathered; together with pest incidence data at the time a stem was harvested.

Two sites were chosen in the area of greatest degradation of the banana shambas, Nyakato to the north of Bukoba and Kibeta to the south of the town.

Using the principle of single stools as the basic experimental unit, or replicate, the trials involved ten treatments each with ten replicates, so requiring 100 stools at each site. These were laid out to give ten rows each of ten stools, with an inter-stool spacing of 3 m and an overall site dimension of 30 x 30m. Apart from Treatment 1, which was planted with 50 cm high sword suckers, the planting material was of the pared corm type; with one sucker or corm per planting hole, as the parent for the stool. It was felt that the difficulty of obtaining sufficient monoclonal planting material could be an additional factor of interest, rather than being detrimental to the outcome of the research.

Two pesticides were used; carbofuran in the form of 5% a.i. granules (Furadan 5C) and isazophos in the form of 10% a.i. granules (Miral 106). An inorganic manure mixture, referred to as NPK mixture, giving a treatment per stool of 296 g calcium nitrate, 110 g triple super-phosphate and 536 g potassium sulphate was used for comparison with cow manure. The measure of cow manure was a common 20 l container known as a debe.

The ten treatments were: (quantities given per stool)

• Treatment 1 - five debes cow manure plus carborufan 3 g a.i.
• Treatment 2 - one debe of cow manure.
• Treatment 3 - five debes of cow manure.
• Treatment 4 - NPK mixture
• Treatment 5 - carbofuran 3 g a.i.
• Treatment 6 - isazophos 3 g a.i.
• Treatment 7 - five debes cow manure plus carbofuran 3 g a.i.
• Treatment 8 - five debes cow manure plus isazophos 3 g a.i.
• Treatment 9 - NPK mixture plus carbofuran 3 g a.i.
• Treatment 10 - NPK mixture plus isazophos 3 g a.i.

The site was laid out following a completely randomized design, using a random number generator to assign treatments per stool. The two trials were very closely supervised and monitored by research personnel.

As FIGURE 1 shows, the sites, assuming all were planted successfully, would have been distributed throughout most of the banana growing areas of the Region, at low, 1200-1300 m, and high, 1600-1700 m, altitudes and also would be grouped by rainfall.

TABLE 1 gives a summary of the potential information. The effect of treatments had the potential to lead to varying nutrition and pest status and this also is given in Table 1. Thus the anticipation was of gaining answers to some at least of the outstanding questions, on aspects such as; crop growth and vigour in different climatic areas; comparison of planting sword suckers or pared corms; pest distribution, incidence and effect on yield (the development and evaluation of crop loss assessment methods was envisaged); the merits of soil amendments, inorganic and organic; the merits of pesticides, carbamate and organo-phosphate types; and, the growth and vigour of different banana clones. Additionally, the involvement of extension personnel and of local farmers in the trials permitted the use of the sites for demonstration and training purposes.

For the factorial trials record sheets were designed by me and printed locally - click for an example.

NEXT

href="\bananas2000\FTRIALPL.HTM"