The scientific, social, and economic significance of the knowledge of boreal forest fire regimes is becoming increasingly evident in North America. One synoptic indicator of fire regimes is the probability distribution of the fire size events, commonly known as the fire size distribution (FSD). For example, information about FSDs is crucial to designing long-term forest management plans that attempt to emulate forest fire regimes. We reviewed the scientific literature to examine the state of knowledge of FSDs with respect to underlying concepts, factors affecting the variability of FSD over time and space, research trends, and knowledge gaps. We found that no single probability distribution addresses all FSDs. There are reports of applications of power law, negative exponential, Pareto, truncated power law, extremal, competing-hazards model distribution types. However, the power-law family, in particular the truncated power law, seems to be most representative of the mid-range of fire eventsizes. Even then, the specific parameters of these probability distributions differ among geo-climate, forest types, and human influences such as fire management strategies. Even in a given space, FSDs may differ depending on the observation period of a fire regime, as well as due to shifts in influencing factors such as climate. Therefore it is essential to understand FSDs in their spatial and temporal context. The knowledge of FSDs in literature has many gaps. Present knowledge of FSDs stems from few scattered locations and mostly the mid-range of fire sizes. The complexities of factors that affect FSDs have not been clearly understood, even in the form of hypotheses, and most of this knowledge is presently in descriptive form. Hitherto exclusive reliance on empirical methods to understand FSDs has many practical limitations, and they may be partially overcome by resorting to simulation modelling methods, especially those that are process-based. Given the limitations of the scientific knowledge, the use of FSDs in practical applications must be judicious, with clear understanding of underlying assumptions and the variability of FSDs over space and time