This thesis investigates how fable picture books represent and construct reality through language and image. To this end, it draws on a social semiotic view of language and other semiotic modes .(Halliday, 1978; Kress and van Leeuwen, 1996), and provides a first systematic account of the .1 1(. discursive practices involved in the representation of reality on structural, intermodal and ideological levels of fable picture books. Using a dataset of six picture books featuring contemporary versions of the Aesopic fable "The Tortoise and the Hare", the present thesis explores similarities and differences in representation by developing a three-level methodological framework. In the first level of analysis, the fable picture books are analysed in terms of their generic structure (Hasan, 1984), which reveals significant links between text and image in the introductory stages of character representation. The second level of analysis expands on Kress and van Leeuwen's (1996, 2006) theory of multimodal communication and explores in more detail the visual representations of characters in the dataset, and makes use of a ,~ system of Balance, as proposed by Painter et al. (2011) to explain composition patters in fable picture books. Finally, the third level unpacks ideological meanings by means of a multimodal Social Actor analysis (van Leeuwen, 2008) and a study of characters' Appearance and Manifestation (painter et aI, 2013). The present study therefore makes an original contribution to a growing body of critical studies on the visual narrative and attests that picture books in .general should be analysed multimodally, due to the equal importance of images and text in the processes of meaning-making. It is hoped that the results of this research will be relevant to teachers using fable picture books in their pedagogical practices, as well as to children's book authors and editors interested in the discourse of fables.