The first step of the Adleman-Lipton extract model in DNA computing is the combinatorial generation of libraries. In this paper a new method is proposed for generating a initial pool, it is a quaternary recombination of strings via application of null context splicing rules. Its implementation, based on a kind of PCR called XPCR, results to be convenient with respect to the standard methods, in terms of efficiency, speed and feasibility. The generation algorithm we propose was tested by a lab experiment here described, since the presence of few sequences is enough for checking the completeness of the library. The simple technology of this approach is interesting in and of itself, and it can have many useful applications in biological contexts.