Maps generated from bird survey data are used for a variety of scientific purposes, but little is known about their bias and precision. We review methods for preparing maps from point count data and appropriate sampling methods for maps based on point counts. Maps based on point counts can be affected by bias associated with incomplete counts, primarily due to changes in proportion counted as a function of observer or habitat differences. Large-scale surveys also generally suffer from regional and temporal variation in sampling intensity. A simulated surface is used to demonstrate sampling principles for maps.