In a blossom, the pistil contains the ova (which have to be fertilized), and the stamen produce the pollen (which will fertilize the ova). Both are found within the same blossom. If isolated (bagged), then a single blossom will fertilize itself--that is, its own pollen will fertilize its own ova.
In deliberate crossing, you establish an isolated female by manually opening the blossom a day or two before it would naturally open, and physically removing the stamen (I believe there are exactly 5 per blossom.), leaving only the pistil in the center. The female-only blossom is then sealed with some paper masking tape. The intended male contributor is obtained by using the pollen from a different variety plant, and transferred on, say, a camel hair brush, to the female-only blossom, which is then resealed until the blossom falls off the developing seed pod.
To produce crosses in both directions (female A and male B, as well as female B and male A), you perform the same approach with each variety. Some of the metabolic determinant genes--contained within the mitochondria--will stay with the female (the plant producing the seed) for each cross.
It might be interesting to do both crosses between two dramatically different variety conformations. For example, crossing in both directions between Big Gem and Little Dutch, might demonstrate if the short short stalk, close node spacing and long, narrow leaves of Little Dutch remain with the female parent Little Dutch. Would "Little Gem" develop on the Little Dutch female, and "Big Dutch" appear on the Big Gem female, or would it not matter much? Making a taller, wider-leafed tobacco with the smoking qualities of Little Dutch would be a boon.
I still worry about the need to have ready blossoms simultaneously on two different varieties. Last season, the blossom time varied as much as a month between the earliest varieties and the latest. This may be mitigated by the long flowering time of some of the varieties.
Bob