Move a diamond ring under a single pendant lamp and it flings tiny points of color across the tablecloth—distinct flecks of red, green, violet, appearing and vanishing as the stone turns. There is no pigment in a diamond to speak of; a gem-grade stone is a block of nearly pure, colorless carbon. Those colors are manufactured on the spot from white light by bending—massive, unequal bending, delivered by one of the most extreme optical materials that occurs naturally. Gemologists call the effect fire, and it is not the same thing as the rainbow sheen of a freshly wiped stainless steel sink: the sink's colors come from thin films interfering with themselves, while a diamond's come from wavelength-dependent refraction, each fleck a genuine little spectrum.
Slowing light down
The first ingredient is retardation. Light entering diamond from air slows to less than half its speed, corresponding to a refractive index near 2.42—among the highest of any transparent gem material; ordinary glass sits near 1.5. High index does two sparkle-founding things. It bends entering light steeply, and it makes total internal reflection easy: inside the stone, light hitting a facet at more than a shallow angle cannot exit at all and rebounds inward. A well-cut diamond is arranged so that light entering the crown tends to bounce twice off pavilion facets and return toward the viewer's eye rather than leaking out the bottom. That returned, reflected volume is the white flash gemologists call brilliance.
Blue takes the wider turn
The second ingredient is the subtle one: diamond's index is not identical for all colors. Tabulated optical data for diamond show its refractive index sliding from roughly 2.45 at the violet end of the visible spectrum to about 2.41 at the deep red end—a difference of a few hundredths, which gemology condenses into a dispersion figure of about 0.044, high for a natural gemstone. Blue light, with the higher index, bends more sharply at every surface; red light bends a little less. Send white light through a facet and the colors do not travel as one beam—they fan out, by amounts that start small and accumulate over two or three internal bounces.
Each facet of a cut diamond is therefore a small prism pointed somewhere in space. When the fanned beam happens to point at your eye, you receive one separated color of that fan for a moment: a single red, green, or violet spark. Turn the stone and different facets re-aim their fans; the sparks migrate, wink out, and reappear. The stone's color comes from where its geometry points rainbow fragments, not from anything tinted inside it.
Cut decides whether it shows
Fire is not a fixed property like weight. Modeling studies of the round brilliant cut, published by researchers at the Gemological Institute of America, traced how brilliance and fire compete for the same incoming light: facet angles and proportions that maximize returned white light are not exactly those that maximize angular spread of color, and every cut trades one against the other. Cut too deep or too shallow and light leaks away before doing either job. This is why two diamonds of identical material can look like different phenomena—one a white searchlight, one a slow drift of colored sparks.
Lighting decides the rest, which is the part of the original observation that rewards notice. Fire is most visible when the ambient light arrives from small, concentrated sources—a candle, a pendant bulb—because each point source becomes a clean, separable fan. Under a large diffuse panel of office light, dozens of overlapping fans average back toward white, and the stone shows mostly brilliance. Jewelers know this; dimmer switches and single bulbs do half the selling.
A third term completes the vocabulary. Scintillation is the pattern of flashes itself—the way the sparkles rearrange as the stone, the light, or you move. The modeling tradition that quantified brilliance and fire treats scintillation as their natural third partner: a static diamond under static light is a fixed arrangement of bright and colored points, and it is motion that converts the arrangement into the live flicker people mean when they say a stone "performs." It also explains the showroom habit of rocking a diamond slightly under a spot: nothing about the gem is changing, but every rock re-aims dozens of tiny prisms at once.
Why lookalikes behave differently
Simulants trip people up in both directions. Glass, with a low index and feeble dispersion, returns little light and splits less color—a poor copy of both effects. Cubic zirconia and moissanite have their own index and dispersion values, different from diamond's, so their balance of white return versus colored spread differs even when the cut is identical. None of this separates simulants reliably by eye at a dinner table—only proper instruments do that—but it does explain why 'sparkly' is not one quality. The rare diamonds that are genuinely colored make the cleanest control experiment of all: their tints come from trace impurities such as nitrogen or boron lodged in the crystal lattice, not from any change to the bending machinery, and their fire plays on top of that body color rather than through it.
If you enjoy sorting closely related hues, the flecks themselves are a tiny exercise in spectral ordering, the same skill a round of Hue Sorting trains at full size.
The two exit rays are labeled with their approximate indices; the wedge shape and ray angles are schematic, so grayscale readers can pair each label with its ray by position.