The Lunatic fringe secretion model is based on experimental data showing the crucial importance of Notch in somite segmentation, on the regulation of L-fng in the PSM, which shows that L-fng is a target of Notch activity (Morales et al., 2002, Barrantes et al., 1999, Dale et al., 2003, Cole et al., 2002), on the action of L-fng on Notch (Blair, 2000), and on the fact that Notch receptor and ligand expression levels seem to be constant in the chick and mouse PSM (Barrantes et al., 1999, Forsberg et al., 1998). It is fundamentally different from previous models of lateral inhibition mediated by Notch signalling (Collier et al., 1996), in that ligand levels are constant, and Notch signalling is periodically active in all cells. It is also very different from the clock and induction model (Schnell and Maini, 2000), in which there is no explicit molecular mechanism, Lunatic fringe is seen as a passive output, and oscillations are cell autonomous (the ratchet mechanism at the heart of that model is made very unlikely by experimental results discussed in section 3.4, which show that a non-zero baseline of Lunatic fringe disrupts oscillations).
The model accounts for coupling between cellular oscillators by secretion of L-fng, which has not been documented in the PSM, but has been in other experimental contexts (Wu et al., 1996). Notch signalling is very versatile and not fully understood, and cleavage of ligands could be involved in signalling (Artavanis-Tsakonas et al., 1999), providing another potential source of coupling; this could require oscillation of Delta levels, as has been observed in zebrafish (Jiang et al., 2000), but would not necessarily be incompatible with Delta levels being constant on a large scale, as in the chick and mouse.
The secreted factor could be different (and involve a more complex mechanism downstream of the clock) without the results presented in this section being necessarily compromised. It is also conceivable that inter-cellular communication could be mediated by gap-junctions (which do exist between PSM cells, Blackshaw and Warner, 1976); in this specific case, the structure of the model could remain the same, but the equations would probably need to be significantly modified to take into account the strong differences between ionic currents and enzyme kinetics.