This article examines the dichotomous relationship between racial hierarchies effected by imperial science, on the one hand, and the subversive potential of the scientific knowledge gleaned from the Andaman Islands, on the other, in Victorian Britain. Knowledge about the Andaman Islands and its ‘savage’ aboriginal tribes had been etched onto British consciousness since the establishment of Britain’s naval base in Greater Andaman (present-day Port Blair), in 1789, followed by a century of anthropological, ethnological, zoological and linguistic and explorations into the Andamanese people. When Sir Arthur Conan Doyle’s The Sign of the Four began being serialised in 1890, a fantastical knowledge of manners and physiognomy of the Andamanese was remarkably familiar to London, through colonial histories, a wide array of photographs in British periodicals, and iconic clay sculptures of the aboriginals displayed at the Colonial and Indian Exhibition in 1886. While British imperialism wanted to project its inexhaustible scientific and technocratic powers counterpoising them against the untameable and (supposedly) prehistoric life of the Andaman Islands, The Sign of the Four ruptured that discourse. I argue that, in the character of the little Andamanese “hell-hound,” Tonga, Doyle presents an example of the failure of imperial scientific prowess to appropriate the savage identity into its racial and hierarchical discourse. Within the seemingly scarce presence of India in the world of Sherlock Holmes, it is deeply consequential that Doyle selected the Andaman Islands as a key location for the origin of his detective plot, as the home of the subaltern Tonga, who pre-empts the spectrality of the hound—a manifestation of imperial guilt and panic—to come later in The Hound of the Baskervilles.