The primary data in this dataset are audio recordings (wav format). They are utterances in the Shilluk language (Nilo-Saharan, South Sudan), recorded from 10 speakers, collected through controlled elicitation. The data were collected in order to investigate the phenomenon of floating quantity in Shilluk, and the analysis of the resulting dataset forms is reported on in the paper "Floating quantity in Shilluk" authored by Remijsen & Ayoker. This paper has been accepted for publication in the journal Language. Apart from the primary data, the dataset includes a segmentation of the vocalic prefix that follows the target word, as described in the above-mentoned paper (in TextGrid objects). In the filenames, the first field holds the speaker code (01-10); the second field specifies the context (utterance-final coded as FIN; utterance-medial as BEN; see Language paper for details); the third field holds the English translation of the target noun (see Language paper for details); the fourth field can be ignored, it is S throughout; the final field holds the repetition number of the realisation, which is almost always 01.