hort communications of practical oriented research/information in agriculture or seed science and technology are presented in this section

Using Scanners to Improve Seed and Seedling Evaluations

Introduction
Seed analysis is a subjective skill relying on the knowledge and expertise of the analyst. As a result, seed testing organizations regularly conduct referee and ring testing programs by which seed analysts evaluate the same seed samples to assure themselves that results are consistent between laboratories. This uniformity of seed testing is vital to ensure the orderly marketing of seeds.  To assist standardization, seed analysts have traditionally relied on educational media such as reference texts, handbooks, manuals, and herbaria. A major limitation of these current identification systems is the inability to share actual seed and seedling specimens with those more skilled in their evaluation. Advances in computer technology offer the promise of (i) easily and cheaply digitizing seed and seedling images using scanners and (ii) rapidly conveying digitized images of seeds and seedlings via e-mail attachments or visits to websites where specimen libraries are maintained. The objective of this study was to develop an imaging platform that could be adapted in a routine seed testing laboratory at little cost to enhance the standardization of purity and germination testing.

Materials and Methods
Seeds: Seeds were placed directly on a flat bed scanner (Hewlett Packard ScanJet 6300C), covered with a green cellophane paper to provide uniform contrast, and captured at maximum resolution (1200 dpi).
Seedlings: Seeds of lettuce (Lactuca sativa L.) and cucumber (Cucumis sativus L.) were germinated on two saturated blue blotters (Anchor Paper Co.) in the lids of 15 x 23 mm plastic boxes. The seeds were vacuum planted in two horizontal rows of 25 seeds each for lettuce, 5 cm each from the top and the bottom of the blotters, and one horizontal row of 15 seeds for cucumber. After planting, the plastic boxes containing the seeds were placed at an angle of 85^o from horizontal in a germinator set at 20^oC for lettuce and 20-30^oC for cucumber.  After 3 days, the plastic lid containing the seedlings was placed on a drawer that was closed under a scanner (UMAX Astra 2000) attached upside-down to the top of a metal box and the images captured at 1200 dpi for photographic reproduction (Fig 1).
Computer: The computer used in this study was a Dell Dimension XPST, 700 MHz, Pentium III with 256 MB SDRAM memory and a 20.4 GB hard drive.

Fig 1. Diagrammatic representation of metal box containing an enclosed scanner fixed upside-down to the lid of the box

Results and Discussion
Images of seeds (Fig 2) could be captured in less than one minute. For seedling images, seeds were germinated in a near vertical position (85^o) on standard blue blotters. This approach provided visibility of the entire shoot and root on a uniform plane against a contrasting color. Germinated seedlings are placed on the drawer of the box, the drawer closed, and the scanner images the seedlings.
        Not all scanners examined functioned upside-down, so it is important to select a scanner that can operate in this position. This approach is superior to inverting seedlings on a traditional flat bed because the seedlings sometimes fall from the blotter or move so that features become obscured.  A further advantage is that root hairs and other delicate structures are not disturbed. It should be emphasized that this scanner possessed a depth of field of at least one centimeter ensuring that all seedling structures remained in focus.

Fig 2. Seeds scanned by HP Scan Jet 6300C flat bed scanner: (A) Radish and  (B) Buffalo grass

Once the seedlings are scanned and digitized into high definition files (1200 dpi), they can be cut and pasted for direct comparison(s) with normal and abnormal seedlings, labeled for structure identification, or saved as compressed JPEG file e-mail attachments for confirmation of normal/abnormal categorization by other seed analysts (Fig 3).

Figure 3. Normal and abnormal lettuce (A) and cucumber (B) seedlings cut and pasted from a germination blotter

Conclusions
This study has demonstrated that flat bed scanners in conjunction with commonly available software to manipulate the digitized seed/seedling images can serve as useful additional tool in a seed testing laboratory.  They possess the following advantages:

• Low cost - each scanner costs less than $300
  
• Simple operation by the seed analyst
  
• Rapid production of digitized images with good depth of field and without lighting
  
• Once captured, the seed and seedling images can be magnified, cut and pasted or sent electronically