The explanation of the Insular Celtic t-preterite as continuing an earlier athematic 3rd sg. s-aorist was not first announced – as is widely held – by Heinrich Zimmer 1890 (‘Keltische studien. 1. Exkurs. Das sogenannte t-präteritum der keltischen sprachen’, Zeitschrift für vergleichende Sprachforschung 30 (1890), 198–217; actually, for Zimmer the starting point was the 3rd sg. middle of the s-aorist), but the idea is already mentioned in a sketch among the notes of Rudolf Thomas Sieg­fried, edited after his death by Whitley Stokes and published 1867 under the title ‘Miscellanea Celtica, by the Late R. T. Siegfried. Collected, Arranged, and Edited by Whitley Stokes’, Transactions of the Philologi‑ cal Society 1867, 252–300 (re-edited in German as: ‘Miscellanea Celtica, von dem verstorbenen R. T. Siegfried. Gesammelt, geordnet und herausgegeben von Whitley Stokes’, vo­l. 6 of Adalbert Kuhn’s Beiträge zur vergleichenden Sprach­forschung auf dem Gebiete der arischen, celtischen und slawischen Sprachen, 1870, 1–18); the entry on p. 296 goes: “S is lost between R and T: Ir. tart = [Goth. thaúrstei], Eng. thirst, TARSTI [Skr. tṛsh. So perhaps the Irish preterite ru-burt ‘tuli,’ from] *ruburṡt, the Welsh kymerth [from *kym-ber-ṡt].”