Some aspects of statistical application in rubber breeding trials


Citation

Tan H., . Some aspects of statistical application in rubber breeding trials. pp. 124-142. ISSN 0127-7065

Abstract

Rubber breeding trials are usually long-term and costly due to the perennial nature of the tree. Thus the experimental economy and precision in rubber breeding trials are important. This paper reviews various aspects of rubber breeding experimental designs applied in two types of trials; small-scale clone trials (SSCT) and large-scale clone trials (LSCT) at the RRIM. In SSCT 6 trials were examined in relation to yield and girth performance using simple lattice design. Average values of trial CV were 8.4 to 26.3 with the highest value for yield per tree (26.3) followed by girth increment (24.0) and girth (8.4 - 10.9). Relative efficiency for simple lattice design over randomised block design ranges from 2.4 to 11.3. In LSCT randomised block design as well as more sophisticated designs (such as simple lattice balanced lattice and incomplete block designs) over randomised block design were also generally low. About 27 of the total cases studied showed more than 20 gain in precision. The findings of 9 -27 of the total cases studied so far having achieved more than 20 gain in efficiency with the more sophisticated experimental designs over the randomised block design in rubber breeding trials merits consideration for the use of sophisticated designs.


Download File

Full text available from:

Abstract

Rubber breeding trials are usually long-term and costly due to the perennial nature of the tree. Thus the experimental economy and precision in rubber breeding trials are important. This paper reviews various aspects of rubber breeding experimental designs applied in two types of trials; small-scale clone trials (SSCT) and large-scale clone trials (LSCT) at the RRIM. In SSCT 6 trials were examined in relation to yield and girth performance using simple lattice design. Average values of trial CV were 8.4 to 26.3 with the highest value for yield per tree (26.3) followed by girth increment (24.0) and girth (8.4 - 10.9). Relative efficiency for simple lattice design over randomised block design ranges from 2.4 to 11.3. In LSCT randomised block design as well as more sophisticated designs (such as simple lattice balanced lattice and incomplete block designs) over randomised block design were also generally low. About 27 of the total cases studied showed more than 20 gain in precision. The findings of 9 -27 of the total cases studied so far having achieved more than 20 gain in efficiency with the more sophisticated experimental designs over the randomised block design in rubber breeding trials merits consideration for the use of sophisticated designs.

Additional Metadata

[error in script]
Item Type: Article
AGROVOC Term: Rubber crops
AGROVOC Term: Hevea brasiliensis
AGROVOC Term: Plant breeding
AGROVOC Term: Trials
AGROVOC Term: Perennial crops
AGROVOC Term: Testing
AGROVOC Term: Tapping
AGROVOC Term: Clones
AGROVOC Term: Crop yield
AGROVOC Term: Replication
Geographical Term: Malaysia
Depositing User: Ms. Suzila Mohamad Kasim
Last Modified: 28 Apr 2025 03:58
URI: http://webagris.upm.edu.my/id/eprint/23387

Actions (login required)

View Item View Item