Rather than using slits in a solid enclosure, vorticity can also be added to the system via arotating screen (Figure 4d ). This setup has been advantageous in that the circulation strengthwithin the domain can be varied through adjustment of the angular velocity of the screen. Emmons& Ying (1967) first adopted this approach, and as a result, it is often called an Emmons-type firewhirl generator. Because the strength of eddies can be controlled, this approach is favorable fortheoretical analysis of the fire whirl structure (Chuah & Kushida 2007), although the domaininstrumentation and measurements are more difficult than in fixed-frame setups. This methodhas also been employed in a series of experiments where multiple equidistant fire whirls weregenerated between two vertical screens that were both parallel to a propane line fire and movingin opposite directions (Lee & Garris 1969). There are other mechanical methods that can leadto the generation of whirling flames, including the use of air curtains and tangentially orientedblowers at the fuel surface (Byram & Martin 1970, Mullen & Maxworthy 1977, Wang et al. 2015).One advantage of these techniques is that restrictive walls are not necessarily needed, which enableseasier experimental probing. However, maintaining the flow symmetry inside the domain is morechallenging than with other methods.While the differences among on-source fire whirls formed in enclosed configurations shownin Figure 4 are minimal, some differences are observed when the fuel source is changed betweena liquid pool fire, a gas burner, or solid combustibles (Hartl 2016, Hartl & Smits 2016). For poolfires and solid combustibles (Martin et al. 1976), the velocity of the gasified vapor is minimal,unlike a gas burner, which may impart additional momentum in the axial direction of the firewhirl, similar to a swirling jet. In a liquid pool fire (Byram & Martin 1962, 1970, Snegirev et al.2004, Chuah et al. 2009, Kuwana et al. 2011), the swirl establishes an Ekman-like boundarylayer over the fuel surface, which draws the flame sheet closer to the fuel surface (Dobashi et al.2015). This provides more heat to the fuel surface and increases the fuel evaporation and burningrate, subsequently entraining more air (this is discussed in more detail in Section 5.3). There aredifficulties characterizing fire whirls over liquid pools, such as the precession of the whirl aroundthe enclosure, uneven heating, and subsequent variations in the heat-release rate. Experimentally,