It has often been presumed that scribes modified the legibility of their text in accordance with the use to which it would be put. The large hands identified in some early Christian manuscripts have been attributed by some to the use of such texts for recitation. The conditions which may have provoked the use of a large script are multiple, as has been recognised. The use of other lectional signs likewise cannot be taken as proof of recitation. Without a source body certainly used for such purposes, we cannot be decisive about the significance of these physical modifications to the text.

The transcription of magical practices onto papyrus would seem to offer an opportunity to observe a scribal attentiveness to recitation. Formulae and voces magicae had to be pronounced exactly to ensure the efficacy of the rituals involved. Therefore the accurate transmission and presentation of such phrases was more significant than in other textual traditions. Nevertheless, the textual variability apparent suggests that this sensitivity did not necessarily result in fewer corruptions to the text. Yet the frequent use of lectional signs and other techniques used to mark out such phrases attests to a concern for legibility. By examining the treatment of formulae and voces magicae within two compilations of magical material, PGM IV and VII, this paper will try to isolate the ritual pronouncements in order to test whether passages for recitation promoted a specific scribal treatment. This work forms part of a wider investigation of scribal practice within the scope of an Australian Research Council project ‘Knowledge transfer and administrative professionalism in a pre-typographic society: Observing the scribe at work in Roman and Early Islamic Egypt’.