I know it isn't quite in the spirit of the article, but what about projecting a negative image into the egg on an enlarger? That'd give you the
in-egg positive.