In this paper we examine the role of multiple emplacement of sills into partly solidified rocks (an intrusive mechanism ‘liquid into solid’) as a possible explanation for some textural and compositional ‘anomalies’ of single-cyclic mafic intrusions. As a case study we used the Shiant Isles Main Sill that is widely regarded as a classical example of a multiple, picrite–picrodolerite–crinanite alkaline sill. This sill is currently interpreted as having been formed by several olivine phenocryst-rich pulses of magma, which were successively emplaced into their almost solidified predecessors. Such an intrusive mechanism is a random process in which many parameters vary independently and unpredictably. Among them are: the number, relative volume and bulk composition of magma pulses, and their place, sequence and timing of emplacement, as well as modal abundance, phase composition and distribution of intratelluric phenocrysts in magmas upon emplacement. In terms of these variables, one can envisage countless different profiles through alkaline sills produced from only three randomly intruded magma pulses of picritic, picrodoleritic and crinanitic composition. Such multiple sills can readily be distinguished from simple ones formed from a single pulse of magma by anomalous compositional profiles with several prominent breaks in crystallization and compositional sequences. The compositional profile of the Shiant Isles Main Sill is remarkably similar to an M-shaped profile expected from fractional crystallization of a single pulse of olivine-saturated magma along a crystallization path Ol+Sp+L (picrite), Ol+Pl±Sp+L (picrodolerite = troctolite), Ol+Pl+Cpx+L (crinanite). The probability of the accidental formation of such a compositional profile by multiple intrusion ‘liquid into solid’ is exceedingly small, even for the single case of the Shiant Isles Main Sill. The probability approaches zero when considering that exactly the same sequence of intrusive events must have been repeated in about 20 neighbouring alkaline sills with similar compositional profiles. This can only be achieved by some universally operating differentiation process. The best candidate for this is the classical fractional crystallization of magma constrained by liquidus phase equilibria. This suggests that the Shiant Isles Main Sill can be best interpreted and modelled as a simple sill that crystallized from one large pulse of magma, with possible involvement of minor refilling events. Further progress in our knowledge of intrachamber magma fractionation processes will probably enable us to interpret many ‘anomalous’ textural and compositional features of mafic–ultramafic intrusions in the frame of a single magma pulse model.