The frequency distribution of stratigraphic layer thickness in cores and a snowmine at South Pole is not compatible with a significant number (>1%) of missing layers associated with zero-accumulation years inferred from pole-height measurements. A reconciliation of these data sets is needed if observed stratigraphic records are to be used as reliable paleoclimate indicators. Three explanations for the discrepancy are offered, namely (i) during a significant number of years, a visible stratigraphic horizon does not form or is not identified, (ii) the true distribution is characterized by two maxima, with a secondary maximum centered around zero layer thickness, or (iii) the pole-height measurements are misinterpreted and there are very few zero-accumulation years at South Pole. With the currently available data, it is not possible to discriminate among these three possibilities. ¿ 1999 American Geophysical Union