Previous studies have explored the size and word type composition (nouns, predicates, etc.) of expressive vocabularies of preschool children with Down syndrome, both spoken and signed. Separately, overall preferences for modality of expression have also been explored. To extend previous findings by describing the relationships between expressive vocabulary size and both word type and modality of expression in the preschool period including changes to modality preference over time. Mothers of 35 children with Down syndrome, aged 36-66 months and attending the same early intervention programme, completed a version - with both spoken word and sign options - of the New Zealand MacArthur-Bates Communicative Development Inventory (CDI): Words and Sentences. Most mothers completed the CDI on multiple occasions across the duration of the study. Completions (n = 114) were analysed cross-sectionally and longitudinally in terms of (1) word type relative to vocabulary size, (2) modality of expression relative to word type and vocabulary size, and (3) individual trajectories in vocabulary size and modality of expression. (1) Word type relative to vocabulary size was similar to previous studies with a greater proportion of the SOCIAL words being present in the children's vocabularies throughout, followed by NOUNS, PREDICATES and CLOSED class words, with proportions converging as vocabulary sizes increase. (2) An initial spoken word preference for SOCIAL and CLOSED class words and sign preference for NOUNS and PREDICATES was found, with more spoken words in larger vocabularies overall. (3) Individual trajectories were highly variable and also revealed temporary points of regression in overall expressive vocabulary size in some children. Children who shifted from reliance on sign to predominantly spoken word expression did so at different ages and at different vocabulary sizes. At school entry, while most of the children used both modalities, some children continued to rely on sign for most vocabulary items whereas others used only spoken words. An appreciation of both the general trends and potential for individual variation in vocabulary structure, modality of expression and change over time will better position clinicians and education specialists to provide individually tailored support to both preschool and school-aged children with Down syndrome. What is already known on this subject Preschool children with Down syndrome have been shown to demonstrate developmental patterns in overall conceptual vocabulary size and word type development broadly similar to typical development. The use of signed vocabulary has also been explored, but independently from word type development and with less attention to individual trajectories in either vocabulary size development or modality preferences throughout the preschool period than is necessary for clinical use. What this paper adds to existing knowledge This study found interactions between word type and modality of expression in relation to vocabulary size as well as a few temporary vocabulary size regressions similar to those found in much younger typically developing children. An initial spoken word preference for social and closed class words, and sign preference for nouns and predicates, was followed by a general trend towards spoken words for all word types as vocabulary (with some fluctuations) increased over time. Children demonstrated considerable individual variation in size of vocabulary and age at which they pivoted from a signed to a spoken word preference. Moreover, while most children used both modalities at school entry, a few retained a sign-only vocabulary and others had moved fully onto spoken word-only expression. What are the potential or actual clinical implications of this work? An understanding of both the general trends and individual variation in vocabulary structure and modality of expression development will better position clinicians and educational specialists to provide individually tailored support to children with Down syndrome.