With plants more than any other living organism, man has been interested in the modifications of characteristics through experiments of cross breeding. Parent plants were carefully selected based on their contributing factors to the next improved generation. It was through such processes that changed the ancestor of corn into the hybrid we see today.
On a genetic level, this all has to do with DNA. If one were to represent each nitrogen base (that make up individual DNA codes) with a letter, it would take 1.7 million pages to contain all the genetic information of corn. If one were to cross two breeds of corn to get an improved plant, genetic laws state that one would end up with not 3.4 million pages but 1.7 million pages. This means that there is random retention of certain parental gene code sequences. In addition, such sequences can be rearranged during the normal fertilization process resulting in different portions of the sequences juxtaposed in offspring. The breeder has no control over which sequences are kept or lost and can only choose offspring with the desired outer characteristics.
In genetic engineering, the process is similar except that instead of the phenomenon occurring inside the cell it takes place in the laboratory. However, whereas the classical breeder cannot read the "language" of those 1.7 million pages of code sequences, the lab scientist can. In contrast to classical breeding, the DNA sources need not be related because they are not sexual crosses. Therefore, asexual plants can be crossed with sexual plants, bacteria, even mammals. In addition, it is possible to control into which tissues the spliced gene is expressed.
Genetic engineering of plants began with fundamental experiments. In one instance, scientists found that a naturally occurring bacterium, Agrobacterium, could infect plant cells by injecting some of its own DNA into plants and having it become a heritable part of the plant cell's genome.