Protocols for in situ hybridization (ISH) of cultured cells often include storage in alcohol at −20°C between fixation of the cultures and the ISH procedure. In experiments aimed at localizing ferritin mRNA in C2 muscle cultures by ISH with digoxigenin-labelled riboprobes, we have noticed that omission of the ethanol storage dramatically changed the pattern of mRNA localization. In cultures stored in 50%, 70%, or 90% ethanol for at least 15 min, ferritin signal was stronger on myotubes than myoblasts but was uniformly distributed over both. In untreated cultures, the signal was patchy, concentrated on the extremities of the elongated myoblasts and very sparse in myotubes. Similar results were obtained with a probe to β-actin used as a control, except that signal was higher in myoblasts in all conditions. When the probes were reduced in size to <100 bases from 561 for ferritin and 1150 for actin, the pattern became uniform, regardless of prehybridization treatment. The patchy pattern disappeared when cells were treated with RNase A following hybridization, suggesting that it is non-specific, despite its absence in cultures hybridized with a sense probe. We conclude that incomplete access of RNA probes can result not only in a reduced ISH signal but also in artefactual patterns of mRNA localization.