Possibilités d'amélioration de l'experimentation sur cacaoyers
The failures often experienced in cacao experimentation have led to through examination of certain points in the experimental designs and techniques of analysing the results. It is a question of: - reducing to the greatest extent the variable factors other than those of which it is desired to measure the effects. - Determining the optimum plot size which can depend on various circumstances. - Bringing into the analysis of variance those factors whose effect can be measured. - Being certain of the validity of the analysis of variance through a proper transformation of the data. For these studies, the existing records of the Cocoa Research Station at Nkoemvone (Cameroun) from 1950 to 1960 have been used. Work was carried out on: - the optimum plot sie and the precision of any trial according to the design used. - The correlation between the yield of neighbouring plots in order to measure the effect on a plot of its neighbours. - The relation between trunk diameter at a given age, or better, the increased in diameter between two ages and the final cumulative yield. The results follow: - regarding plot size, two situations must be considered: 1) Where guard rows are not used, single tree plots can be used. 2) Where guard rows are necessary a plot size of 12 - 25 trees is recommended, 25 trees would be chosen when the area is not limited, staff and money are adequate. If funds are limiting, the number of replications may be reduced, but at the expense of the precision of the experiment. If the area is limited, 12 tree plots would be used. In choosing an experimental design, the standard method should be used. - The residual error can be reduced in the analysis by using additional measurements such as: the number and date of loss of tree in each plot, measurements of the trunk at 20 cm from the ground taken at one or two years of age before this factor is influenced by the treatments, if this is dome, the analysis of these measurements allows a rapid sorting out of the material in the experiment
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | biblioteca |
| Published: |
St. Augustine (Trinidad y Tobago)
1972
|
| Subjects: | THEOBROMA CACAO, EXPERIMENTOS, TAMAÑO DE LAS PARCELAS, ANALISIS ESTADISTICO, |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The failures often experienced in cacao experimentation have led to through examination of certain points in the experimental designs and techniques of analysing the results. It is a question of: - reducing to the greatest extent the variable factors other than those of which it is desired to measure the effects. - Determining the optimum plot size which can depend on various circumstances. - Bringing into the analysis of variance those factors whose effect can be measured. - Being certain of the validity of the analysis of variance through a proper transformation of the data. For these studies, the existing records of the Cocoa Research Station at Nkoemvone (Cameroun) from 1950 to 1960 have been used. Work was carried out on: - the optimum plot sie and the precision of any trial according to the design used. - The correlation between the yield of neighbouring plots in order to measure the effect on a plot of its neighbours. - The relation between trunk diameter at a given age, or better, the increased in diameter between two ages and the final cumulative yield. The results follow: - regarding plot size, two situations must be considered: 1) Where guard rows are not used, single tree plots can be used. 2) Where guard rows are necessary a plot size of 12 - 25 trees is recommended, 25 trees would be chosen when the area is not limited, staff and money are adequate. If funds are limiting, the number of replications may be reduced, but at the expense of the precision of the experiment. If the area is limited, 12 tree plots would be used. In choosing an experimental design, the standard method should be used. - The residual error can be reduced in the analysis by using additional measurements such as: the number and date of loss of tree in each plot, measurements of the trunk at 20 cm from the ground taken at one or two years of age before this factor is influenced by the treatments, if this is dome, the analysis of these measurements allows a rapid sorting out of the material in the experiment |
|---|