PROGRESS IN THE DEVELOPMENT OF SAMPLING METHODS TO ESTIMATE COTTON LINT STICKINESS DUE TO SWEETPOTATO WHITEFLY INFESTATION

Authors: Naranjo, Steven; Henneberry, Thomas; Chu, Chang Chi

Submitted to: National Cotton Council Beltwide Cotton Conference
Publication Type: Proceedings
Publication Acceptance Date: 6/1/1998
Publication Date: N/A
Citation: N/A

Interpretive Summary: Cotton stickiness, due primarily to Bemisia, has become a limiting factor in cotton production in many countries, and may presently be considered by the cotton industry as the most serious factor affecting cotton quality. Enzyme-based technologies for reducing or eliminating lint stickiness at either the pre-harvest or post-harvest stage are being developed. To most efficiently deploy these systems it will be necessary to determine whether a field is in need of remedial treatment for stickiness. Standardized laboratory techniques (e.g. thermodetector) are available for indexing the stickiness of lint, but there is no standard methodology for collecting samples from the fields. Research was conducted in 1995-1997 in central AZ and Imperial Valley, CA to examine the distribution of sticky cotton lint, optimize the sample unit size, and determine the number of samples needed for the precise estimation of lint stickiness. Preliminary results indicate ethat lint stickiness is randomly distributed in the field and that the highest cost efficiency is attained with a relatively small sample unit (all the lint from a single plant). Further analysis indicate that each sample from the field should be assayed twice in the laboratory to reduce the variability associated with the thermodetector technique. Preliminary sampling plans have been developed for a 1-plant sample unit.

Technical Abstract: Cotton stickiness, due primarily to Bemisia, has become a limiting factor in cotton production in many countries, and may presently be considered by the cotton industry as the most serious factor affecting cotton quality. Research was conducted in 1995-1997 in central AZ and Imperial Valley, CA to examine the distribution of sticky cotton lint, optimize the sample unit tsize, and determine the number of samples needed for the precise estimatio of lint stickiness. There was no statistical difference in mean estimates of stickiness (thermodetector spots) among 14 different sample units tested. Standard deviations were equal to or less than the mean for all sample units indicating a random or Poisson sampling distribution. Relative net precision declined with increasing size of the sample unit and a 1- plant sample was the most cost-efficient. Partitioning of variance components indicated that a large amount of variability was associated with hreplicate assays on the same sample with little variation in sample-to- sample differences in the field. This analysis suggests that the optimal number of replicate samples to perform in the laboratory will depend on the ratio of field to laboratory costs. With present technology a minimum of two assays should be completed on each sample. Preliminary sampling plans have been developed for the 1-plant sample unit and results suggest that high precision (SE/mean ratio = 0.10) may be achieved with less than 25 sample units.