Shape measurements in horticultural research have generally been expressed as ratios or indexes. Computer-based image analysis enables the objective quantification and statistical analysis of two-dimensional sample shape variability. In addition, the availability of public domain software facilitates the inexpensive but accurate quantification of object shape in horticultural research. We describe the procedures for measuring sample shape using the following publicly available software: ImageJ, ImageTool, and SHAPE. Using U.S. #1 sweetpotato storage root samples from plots subjected to various weed control treatments, we detected significant differences in elongation, compactness, as well as shape attributes. We also measured size and shape variability from representative fruit, leaf, and floral organ samples. The results demonstrate that, where possible, measurement of two-dimensional samples can be undertaken inexpensively and accurately using public domain software applications.
Arthur Villordon and Jason Franklin
Arthur Villordon*, Jason Franklin, and Don LaBonte
The use of handheld computers such as personal digital assistants (PDAs) represents a feasible method of automating the transfer of files to computers for archiving and statistical analysis. Data collected using the PDA can be transferred directly to a database program on a desktop computer, virtually eliminating errors associated with the reentry of manually collected data. These devices are highly portable and can be housed in protective cases, enabling data collection even in inclement environments. The availability of handheld database programs that permit the development of electronic forms further makes the PDA a viable data collection platform for scientific research. These database applications not only allow novice users to develop customized forms that facilitate the recording of alphanumeric data; these applications also synchronize directly with current desktop-based database and spread-sheet applications. We used Microsoft Access database tables, along with Visual CE, a PocketPC database application, to generate electronic forms for collecting data from research trials conducted in 2003. To facilitate comparison with manual data collection, we also recorded observations using “pen and paper” methods. We found no differences between both methods in the length of time required to enter observations. However, the PDA transferred the data to a computer 600% faster relative to the manual reentry method. Using the handheld computer, field data was immediately available for compilation and statistical analysis within minutes of completing the data gathering process, at the same time ensuring the integrity and continuity of the files.
Arthur Villordon, Jason Franklin, and Don LaBonte
Handheld computing devices, such as personal digital assistants (PDAs), can potentially reduce repetitive tasks that pervade data collection activities in horticultural research. PDA-collected records are electronically transferred to a desktop computer, eliminating manual reentry as well as the need of reviewing for incorrect data entries. In addition, PDAs can be enclosed in protective cases, enabling data collection in inclement weather. Visual CE-generated database forms installed on PDAs were used to electronically collect data from research trials conducted in 2003. The records were subsequently transferred to Microsoft Access desktop database tables for archiving and subsequent statistical analyses. Data for certain trials were also manually collected using paper forms to facilitate comparison between manual and PDA-assisted data collection methods under controlled conditions. Using paired samples analysis, we determined that electronic transfer of records reduced the time required to store the records into desktop computer files. Manual and PDA-based recording methods did not vary in the time required to enter numerical measurements. Our experience demonstrates that off-the-shelf software and consumer PDA devices are viable options for data collection in research. PDA-assisted data collection is potentially useful in situations where remote, site-specific records need to be merged into a central database and where standardized measurements and observations are essential for performing analysis.