Preparation of a sample of beewings is described in the Bibba publication "Breeding Better Bees using simple modern methods".
Bibba recommend a sample size of thirty wings (not fifteen pairs). This program and spreadsheet will work with any sample size from 1 to 50 wings.
To use this program it is essential that all wings are presented as right wings viewed from above the bee. If you choose to use left wings then they need to scanned from the other side or the scans manipulated to create a mirror image.
If you are using a flat bed scanner and are not concerned about saving the sample a minimalist approach to sample preparation can be used. In draught free conditions arrange your sample of bee wings on the bed of the scanner, all the right way up (ie so the scanner will see right wings) and close together, carefully cover them with a piece of clean white card, and scan them. It is not essential for them to be arranged in neat rows but it helps to keep track of the wings when you are clicking on each wing in turn. If they are close together then the scan size is minimized so that it is possible to do more wings with each scan.
Scans should be made at a resolution of at least 2400 dots per inch even if your scanner doesn't support this high a resolution. This is necessary for consistent results. Your technique in clicking on each point will affect consistency too.
There are several potential sources of error in this process.
1. The mouse pointer cannot be positioned sufficiently precisely. Scanning at a higher resolution, even if this is above the scanner's maximum optical resolution, can minimize this.
2. It may be difficult to see precisely where you should position the mouse pointer, in this case a scanner with a higher optical resolution may show more detail.
3. The scanner needs to be free from distortion and consistent in horizontal and vertical directions.
Files should be saved as bitmaps (.bmp) or jpegs (.jpg) to work with this program. Jpegs have the advantage that they take less disc space.
Bibba advise that when using a projector method measurements should be made to an accuracy of half a millimetre at 40 times magnification. This is equivalent to making a direct measurement on the bee's wing to an acuracy of 0.0005inches (0.0125mm). On a scan at 1200 dots per inch (dpi) this is equivalent to a permitted error of 0.6 of a pixel.
One of the limitations of a scanned picture is that we can only select whole pixels. If you are sufficiently skilled with your mouse you will be able to select the pixel containing the correct endpoint of a line but as the line will effectively be measured to the centre of the pixel then there will be an average error of half a pixel (0.25 at each end of a line). This is only just within the permitted error calculated in the previous paragraph. It follows then that 1200 dpi is the minimum resolution for scanning beewings. Greater accuracy can be obtained by scanning at higher resolutions. At 2400dpi the maximum error should less than the 0.0005 inches calculated.
The scan resolution should be chosen to provide sufficient accuracy and to make the size of the scanned image manageable. 2400 dpi seems sufficient for consistency. The total size of the scan can be minimized when preparing the sample by keeping the sample wings close together. This program has been tested with images up to 7200 x 3600 pixels (3 X 1.5 iches at 2400dpi). Note that the sample of 30 (or more) wings can be prepared as more than one scan if necessary.
As discussed above one of the main advantages of high resolution is the ability to position the cursor more precisely. This will still be the case if you use an interpolated resolution (ie where the scanner software works out the intermediate pixels) so a scanner with 600 dpi optical resolution can still produce consistent results.
A scanner can be tested for distortion by scanning a square, turning it through 90 degrees to rescan it and comparing the two scans.