This bipartite thesis presents and implements an intermedial model for co-reading poems and photographs in paper books, a genre I call photo poetry. A survey of the genre in Sweden was carried out and presented in a selected bibliography in my licentiate thesis. Two well-established poets, Katarina Frostenson and Rut Hillarp, made their debut in the genre in the 1980s and have since produced three books each. Frostenson cooperates with photographer Jean Claude Arnault and Hillarp created the poems as well as the pictures herself. Scholarly studies up to now have focused on the poems, however, and have therefore neglected the impact of the photographic pictures. The first part of the thesis elaborates a model based on the framework of the picture book and adapted to the text genre of poetry and the epistemology of the photographic picture. Two different narrative reading strategies are developed and applied to the material: the metonymical and the metaphorical linking. The metonymical linking implies that the diegesis on the spread is perceived to be part of a larger diegesis and that that diegesis has direct virtual contact with the diegesis on the next spread. The metaphorical linking implies that the diegesis on the single spread is perceived to be part of a larger diegesis, but that that diegesis has no direct virtual spatial contact with the next spread. Whether or not a diegesis is perceived to have direct virtual spatial contact with spreads depends on the story’s display of the contingency of characters, time and place. The model is based on the relations tied to the book’s construction: the schematic, the synchronic and the diachronic relation. The schematic relation concerns meaning created on all spreads, the synchronic relation meaning on a single spread, and the diachronic relation meanings on spreads in a row. The schematic and the diachronic may appear to overlap somewhat but in the schematic relation the focus is on tracing different story schemes, and in the diachronic relation the focus is on how the narration progresses and alternates between different schemes. The findings show that with the co-reading model the impact of the photographs gives a deeper understanding of not only the narrative interplay of word and image but also of semiotic, intermedial and intertextual connections. The reading strategies applied show that Frostenson’s and Arnault’s works gain from a metonymically linked interpretation whereas Hillarp’s mainly gain from a metaphorically linked interpretation. The study also discusses the impact of the photographic picture and connects it to the semiotic theory of C. S. Peirce as well as to Western picture practices.