The cancer stem cell hypothesis has recently re-emerged as a compelling paradigm for the development and progression of neoplastic disease. The hypothesis proposes that a specific subset of “cancer stem cells” (CSC), believed to share many features with normal stem cells, is exclusively responsible for maintaining tumour growth and driving progression. If the CSC hypothesis applies, it may require re-evaluation of the clinical approach to neoplasia. Spontaneous cancer in the domestic dog represents a significant welfare problem, with dogs developing many tumours strongly reminiscent of those affecting humans. This study sought to investigate whether cells with characteristics of CSC are identifiable in canine cancer. Assays to identify, isolate and characterise CSC were adapted to the canine system, and cancer cell lines and spontaneous tumours of diverse origin evaluated for the presence of candidate populations. Whilst analysis of surface expression patterns did not identify specific subpopulations within canine cancer cell lines, these were detectable in cells derived directly from primary tumours. Assays for stem cellassociated drug resistance mechanisms could also be used to identify subsets of putative canine CSC. Formation of “tumourspheres” by canine cancer cell lines was found to be highly density-dependent, so a potentially unreliable method of isolating CSC. Expression of the cell surface glycoprotein CD44 was associated with cellular proliferation status, although it may not represent a stable canine CSC marker. The NFκB survival pathway, associated with apoptosis resistance of some putative CSC, was constitutively active in canine cancer cell lines; suppression using specific inhibitors could reduce cell viability, indicating that this may represent a rational therapeutic target. Overall, these studies demonstrated that CSC assays may be adapted to the canine model system, although they require rigorous interrogation to distinguish apparent CSC attributes from basic biological properties. Cell lines have provided a stable background upon which to optimise assays, but appear less likely to demonstrate discrete CSC subpopulations. Putative CSC subsets may be more readily identifiable within heterogeneous primary tumour cells. The application of some of these adapted assays within a clinical setting may enable further characterisation of individual patients’ tumours, and inform therapeutic regimes for improved treatment outcomes.