Image Reduction in Photoshop
When you resize an image in Photoshop, it is resampled to the new size, usually with the 'bicubic' algorithm. To calculate the new value of every pixel in the new image, it averages the color values of the corresponding areas of the original image, which results in a general softening.
If you reduce a large scan, say 1000 pixels wide down to 100 pixels wide (10%) the softening can be quite severe and no amount of sharpening can ever restore the image information that has been discarded.
Instead of doing the reduction as a single 'reduce to 10%' in one jump, try this:-
First, reduce the image to 50% and apply the Unsharp Mask filter with values of – Amount = 50%, Radius = 1 pixel, Threshold = 0.
Then do the same again and you will have a sharp image 250 pixels wide.
Finally, reduce to 100 pixels wide (or 40%) and apply the same Unsharp Mask filter.
By reducing and sharpening in steps like this, you will end up with a much crisper image than if you do it in one step. 50% reductions combined with 50% unsharp masking works well for most images but you could try varying the Unsharp Mask settings slightly on the final reduction.