The earliest phase of Arakan history, between about the fifth and the tenth centuries, has to be written on the basis of inscriptions and related material such as coins bearing Sanskrit texts, as well as sculpture and architecture. These show Arakan to have had strong ties to Southeastern Bengal (the Samataṭa and Harikela regions) and beyond this with the Buddhist communities of Northeastern India using Sanskrit as preferential medium of expression. A first batch of Arakan Sanskrit inscriptions was studied by E.H. Johnston and published posthumously in 1943. Since then, this field has moved forward thanks mainly to three articles by D. C. Sircar that appeared in the 1960s and 1970s. Since then, no further epigraphic material of early Arakan has been published. The Sanskrit inscriptions of Arakan are often in deplorable state of preservation. But even fragmentary material can throw new light on the past, especially when studied in combination with epigraphical and numismatic discoveries made in Southeast Bengal over the past half-century. This article deals with three such fragmentary inscriptions, all previously unpublished. It presents the discovery that the ancient name of Arakan was Kāmaraṅga, discusses aspects of the history of Buddhism in Arakan during the first millennium, and discusses the problem of the chronology of early Arakan based on a detailed palaeographic analysis of the inscriptions published here.