In June 1980 Turner Whitted, then at Bell Laboratories in Holmdel, New Jersey, published “An Improved Illumination Model for Shaded Display” in Communications of the ACM, volume 23, number 6, pages 343 to 349. The paper is the foundational description of recursive ray tracing as a complete rendering technique. Its central observation is that to shade a point accurately you need global information about the rest of the scene, and that this information can be gathered by following rays of light. The work appeared first as a SIGGRAPH 1980 presentation and was published in the journal the same year.
Whitted’s model replaces the local shading calculation of earlier methods with a tree of rays grown for each pixel. A primary ray is cast from the eye through the pixel into the scene. Where it strikes a surface, the shader may spawn additional rays: a reflection ray in the mirror direction, a refraction ray bent according to Snell’s law for transparent materials, and shadow rays toward each light source to test whether the point is occluded. Each spawned ray can in turn hit another surface and spawn more rays, so the computation is naturally recursive and terminates when rays escape the scene, reach a light, or fall below a contribution threshold.
The result was the first rendering method to produce sharp mirror reflections, correct refraction through glass and water, and accurate shadows from a single unified algorithm rather than from separate special-case tricks. The paper’s images of reflective and transparent spheres above a checkerboard became among the most recognizable pictures in computer graphics. Whitted also contributed practical engineering: an efficient ray and bounding-sphere intersection scheme that made the recursion affordable on the hardware of the time.
Whitted built directly on Arthur Appel’s 1968 idea of casting rays to determine visibility and shading, but extended it from a single bounce to full recursion, which is why the technique is often called Whitted ray tracing or classical ray tracing. The model captures specular interreflection but, by design, does not solve the full transport of diffuse indirect light; that broader problem was later framed by the rendering equation and addressed by radiosity and path tracing.
The paper remains a primary reference because nearly every modern offline renderer, and the real-time ray tracing hardware that shipped decades later, traces its lineage to the recursive ray tree Whitted described. The canonical record is the Communications of the ACM article at DOI 10.1145/358876.358882; the SIGGRAPH reprint distributed by the ACM Digital Library preserves the same text and figures.