The separation of map projections into sets according to criteria uniquely characterising each set. Map projections are commonly classified either by the geometric quality preserved (i.e., left unchanged) by the transformation or by the kind of surface onto which the ellipsoid is mapped as an intermediate step in mapping onto the plane. The geometric qualities used are distances between points, angles, and areas within closed curves (or distances and areas scaled to the size of the map produced. ) A map projection keeping distances from a particular point or line unchanged except for scale is called an equidistant map projection; one keeping the angles between lines unchanged is called a conformal map projection; one keeping the azimuths from a particular point unchanged is called an azimuthal map projection; and one keeping areas within closed figures unchanged except for scale is called an equal area map projection. There are no generally accepted names for other sets. Map projections usually map the ellipsoid either directly onto the plane or first onto a sphere, cone, or cylinder and then from that surface onto the plane. If the map projection is from a cone or cylinder, the mapping onto the plane is done by the simple algebraic analogue of unrolling the cone or cylinder. A spherical surface cannot be unrolled, but the algebraic transformations from ellipsoid to sphere and from sphere to plane are usually very much simpler than the single step process from ellipsoid to plane. A map projection using the cone as an intermediate surface is called a conical map projection; one using the cylinder is called a cylindrical map projection. There is no particular name for the set of map projections using the sphere as an intermediate surface. Except for the aposphere, which has been used as an inter-mediate surface for the transverse Mercator map projection, choice of intermediate surfaces has been pretty much limited to the sphere, the cone, and the cylinder. Map projections are also classified into two sets according as they are true projections (i.e., are equivalent to drawing straight lines from a given point, through the points on the ellipsoid, and onto the plane or intermediate surface) or are not. Those which are true projections are then further classified according to where the centre of projection and the ellipsoid are placed with respect to the plane or intermediate surface. Many other schemes of classification exist but are comparatively little used. The best known are those of Maurer (1935) and Tobler (1962); there are more recent ones by Wray and Chovitz.