The National Serials Pilot Project, Phase II of the National Serials Data Program, is described. Utilizing the MARC format for processing serials, the objectives were: (1) to create a machine-readable file containing live serials in the fields of science and technology, (2) to produce a number of preliminary listings, and (3) to produce one or more written reports covering procedures, problems and results. Data were input via an administrative terminal system to a 360/40 computer; processing of data was done on a 360/50 computer. Among the conclusions and recommendations are: (1) a national serials data bank in machine-readable form is both technically and economically feasible; (2) such a data bank should have its own machine-readable authority file for corporate names; (3) input and output in upper case only would be more satisfactory from both the systems viewpoint and the cost viewpoint, but probably would not be accepted by the library community; and (4) serious consideration should be given to the question of applicability of existing cataloging rules in the determination of main entry in a machine-readable file. (Author)