Purpose: The purpose of this paper is to report an investigation into the types of knowledge organisation systems (KOSs) utilised in North American digital library collections. Design/methodology/approach: The paper identifies, analyses and deep scans online North American hosted digital libraries. It reviews the literature related to the application of KOSs on the web, identifies widely used KOSs and tools and reviews the literature related to collaborative collections on the web. Findings: A total of 269 North American digital library collections were examined. The Library of Congress Subject Headings is the most widely used subject representation tool, followed by domain-specific thesauri, 113 digital library collections make use of locally developed taxonomies. A few collections use the Dewy Decimal Classification and alphabetical indexes. Research limitations/implications: This research was limited to North American digital library collections. Practical implications: The findings show the popular KOSs used in digital library collections. It also shows the organisational contexts of the examined digital library collections. Originality/value: This research contributes to the areas of digital libraries and to the application of KOSs and services for subject representation and access.