The proposed model of the genesis of De Geer moraines in Finland is based mainly on the distribution of the moraine areas in relation to the palaeohydrological environment, and on the composition and two- and three-dimensional fabrics of the material. In order to shed some light on the geochronological problem involved, the distances between successive ridges in a number of De Geer moraine series are also compared with clay varve data on deglaciation rates in Finland. Granulometrically, petrographically and morphometrically, De Geer moraine till does not differ significantly from tills of surrounding areas. Two-dimensional fabrics stress orientations at right angles to the ridge crest, and three-dimensional analyses reveal the slope-conformable character of pebble inclinations. Generally, spacings of the De Geer moraine ridges do not confirm annual deglaciation rates as conveyed by clay varve data. It is suggested the moraines were formed in basal crevasses of the ice, following active flow phases of a local surging nature. During the quiescent phase of the surge cycle, the ice subsided into a water-soaked till mattress, filling up basal cavities from the proximal or the distal side. Finnish De Geer moraines are thus neither annual nor end moraines. They are subglacial bedforms with no geochronological meaning, except that they formed during one single surge cycle. As the marginal zone of the ice sheet calved and thinned, it was lifted by proglacial water and the whole group of moraines became free of ice almost simultaneously.