About the Craster Projection
J.E.E. Craster experimented with different equal area pseudo-cylindrical projections in the 1920s. These used hyperbolas and ellipses for the meridians, but he eventually settled on parabolic meridians in 1929. A number of textbooks used the Craster Parabolic projection in the 1950s. Notably, an oblique form (ie. with an unconventional pole and central parallel) of the projection was used by the National Geographic Society in 1971 for a map of Asia.
The Craster Parabolic projection is a pseudo-cylindrical projection. Ie. a mathematical projection with straight line parallels (like a cylindrical projection) but with curved meridians (unlike a cylindrical projection). With the exception of the central meridian, the meridians are all parabolas. The end result is a projection which resembles the Sinusoidal projection but it is slightly more rounded.
Author's Note: Of the projections presented here at Equal-Area-Maps.com, the Craster projection is probably best classed as one of the "also rans". It has moderate software support, with support from both ESRI and Proj4, but not Proj4JS. Although it is slightly more aesthetically pleasing than the Sinusoidal projection, the Mollweide projection is better than both. These latter two projections also have better software support.
See our interactive Craster Projection world map (overlays disabled due to the lack of Proj4JS support).