Allan M. Maxam and Walter Gilbert, PNAS 74(2), 560–564 (1977), DOI doi.org/10.1073/pnas.74.2.560 . The primary abstract describes terminal labeling, base-selective partial cleavage and size separation of labeled fragments. The original first-page full byline ALLAN M. MAXAM AND WALTER GILBERT is verified in the university-hosted paper scan hackert.cm.utexas.edu/courses/bch370/fall2014/StudentArticles/Maxam%20Gilbert%20Seq%20PNAS%201977.pdf .
Additional verified original-paper locators: printed pp. 560–561, An Example, used a 64-base-pair lac-operon fragment; the authors read 62 bases of each strand. Printed p. 561 explicitly states that the fragments end with the base immediately before the base destroyed by chemical attack. The original purine displays are strong-G/weak-A and adenine-enhanced, rather than assuming four exclusive base reactions. hackert.cm.utexas.edu/courses/bch370/fall2014/StudentArticles/Maxam%20Gilbert%20Seq%20PNAS%201977.pdf
Walter Gilbert and Benno Müller-Hill, Proceedings of the National Academy of Sciences USA 56(6), 1891–1898, December 1966. DOI 10.1073/pnas.56.6.1891. The original first-page byline and opening negative-control/operator-binding model are also indexed in the university-hosted scan: www.bio.davidson.edu/courses/synthetic/papers/Gibert_Muller_hill_66.pdf .
Additional exact primary locator: printed p. 1891, assay description and Design of the Experiment, in the already listed university-hosted original scan www.bio.davidson.edu/courses/synthetic/papers/Gibert_Muller_hill_66.pdf . The assay detects binding to inducer by equilibrium dialysis against radioactive IPTG, deliberately avoiding dependence on the exact repression mechanism. This is distinct from the opening operator-DNA-binding model and adds experimental specificity to the existing source card.