Abstract An investigation of a landslide involving a federal highway, the Umuahia—Bende Road, in southeastern Nigeria is presented. The slide occurred in an apparently stabilized slope one year after a previous “corrected” slide. Laboratory tests show that residual strength parameters as obtained in multiple reversal direct shear tests could be used reliably in the prediction of fresh slides in pre-existing but apparently stabilized slides, provided that zero cohesion is assumed even if laboratory results indicate the existence of the same. It is believed that cohesion is progressively lost with displacement and time and that laboratory tests do not simulate enough of the time period of the cohesion loss in the field. The study shows that a wrongly “corrected” slide could remain quasi-stable for as long as one year but would become unstable again under the effects of adverse conditions such as rise in water table.