adult organ and grown in culture under different conditions. In 2D cultures at an air– liquid interface or in a microfluidics (lung-on-a-chip) device, airway stem cells self-organise into a pseudostratified mucociliary epithelium. In 3D cultures, stem cells form structures known as organoids, which differ in organisation depending on which population they are derived from [6–8]. These ex vivo culture systems have tremendous potential for modelling pulmonary disease and for drug screening. Before we summarise the different stem cells found in the adult lung, some historical background is in order. Based on early studies, it was assumed that stem cells must be unspecialised and quiescent. In fact, it is now recognised that some stem cells, such as those in the crypts of the small intestine, normally proliferate quite actively, while others have specialised physiological functions [9]. For example, in the case of the lung, the type 2 alveolar stem cells, which can self-renew and differentiate into type 1 cells, are specialised to secrete surfactants, and to recruit and activate immune cells [10], while the myoepithelial stem cells of the SMGs are contractile and express smooth muscle actin. It was also assumed that undifferentiated “professional” stem cells would have a deterministic pattern of behaviour, giving rise after division to either two stem cells (symmetric behaviour) or to one stem cell and one differentiating daughter (asymmetric behaviour). It is now clear that not all stem cells follow these rules, and that alternative models are possible, even for stem cells in different regions of the same organ. Thus, in some cases, stem cells are best viewed as a heterogeneous population of cells with varying probabilities of giving rise to either two stem cells, two differentiating daughters, or one of each. Cells can transition reversibly between these states, and the probability of each decision can vary depending on the intrinsic state of the cell and signals from the local microenvironment [11, 12]. These different models mean that any new prospective stem cell type must be studied quantitatively over both the short and long term, ideally using lineage tracing, live imaging and single cell transcriptomic methods under different physiological conditions [12–14]. Another feature of adult stem cells that has emerged from recent studies is the fact that the fate of the differentiating daughter cells is not invariant but can change with signals from the environment. Abnormal conditions can also trigger some differentiated cells to “dedifferentiate” or “transdifferentiate” back into stem cells [4, 15, 16]. This “cell plasticity” is particularly evident in response to injury or inflammation. Such conditions are frequently encountered in studies on the adult lung because cell turnover is normally very slow in order for the full potential of stem cells to be realised, or for new reserves to be revealed, it is necessary to experimentally damage the tissue and to follow repair and remodelling over time. As described in several chapters, a wide range of injury/repair models are typically used in the mouse lung, most of them affecting the epithelium. They include exposure to detergent (polidocanol), acid (or sulfur dioxide), ozone, naphthalene, elastase and bleomycin, as well as virus infection and cell-specific conditional deletion using diphtheria toxin. What these studies have revealed is that the identity and fate of activated stem cells can vary depending on the severity and nature of the injury and the age of the animal. Therefore, to understand the in vivo functional role of any stem cell population, it is important to use as many different experimental variables as possible, and to follow the fate of stem cell descendants quantitatively over long periods of time. Figure 1 summarises the epithelial/endothelial stem cells that have been identified to date in the adult mouse lung as yet, very little is known about the lineage relationship among adult mesenchymal cells. The figure indicates whether similar stem cells have been identified in the human lung, but here our knowledge base is also very limited, and x https://doi.org/10.1183/2312508X.10002321