Calculate the Color Index of an Igneous Rock
The color index is the volume percentage of dark, ferromagnesian minerals — biotite, amphibole, pyroxene, olivine — relative to the total volume of the rock. It is a classification criterion for igneous rocks in its own right, independent of their silica content.
Counted minerals (points or area, as for modal composition) are listed with each phase tagged dark or light: the calculator returns the color index and the rock's class.
Common classification: leucocratic (CI < 35%), mesocratic (35–65%), melanocratic (65–90%), ultramafic (CI > 90%).
Counted minerals, dark and light
Technical labels followed by * are listed in the glossary.
Felsic versus mafic: the reference mineral lists
The color index* (CI, sometimes M or M′) is a volume percentage, not a paint chart. Every mineral is sorted dark or light, then divided.
- Light ( felsic*). Quartz. Alkali feldspar (orthoclase, microcline, perthite). Plagioclase, even the grey calcic ones — they count as light. Muscovite, when present. Feldspathoids (nepheline, leucite) go here too: pale, no Fe–Mg to speak of.
- Dark ( ferromagnesian*, mafic). Olivine. Pyroxene (augite, hypersthene, diopside). Amphibole (hornblende). Biotite. Opaque oxides — magnetite, ilmenite — unless the lab has a written rule that throws them out. Chlorite after biotite still sits on the dark side of the ledger; the rock did not become leucocratic by weathering green.
CI = Σ vol.(dark minerals) / vol.(rock) × 100
A point count on a thin section is the same count already done for modes; each phase is simply tagged dark or light. A mineral on the fence (some epidote, some tourmaline) needs a house rule kept for the whole suite. Some authors reserve “mafic color index” for olivine + pyroxene + amphibole + biotite + opaques and keep muscovite or epidote out; IUGS M′ is close to that list in ordinary plutonic rocks. One definition per paper. This calculator takes whatever is tagged dark.
Color index as a complement to silica content
SiO2 from a whole-rock analysis is the chemical axis: felsic > ~63%, intermediate ~52–63%, mafic ~45–52%, ultramafic below that. There is no XRF on the outcrop. There is a broken face and a hand lens.
Color index tracks silica because Fe–Mg minerals are the silica-poor ones — most of the time. A leucocratic granite and a rhyolite sit at low CI and high SiO2; a gabbro and a basalt sit at high CI and low SiO2. The overlap is the point of having both numbers: an altered, bleached mafic rock can look pale and still be low-silica; a biotite-rich granite can look salt-and-pepper and still assay 70% SiO2. CI is a field and thin-section axis. Silica is a geochemistry axis. Together, not as synonyms.
Leucocratic to ultramafic: what each class looks like on a fresh face
IUGS / Streckeisen cuts, the ones this calculator uses: 35, 65 and 90. Not 30 and 60. The thresholds belong on the notebook page; a later mapper should not “correct” them.
- Leucocratic (CI < 35%) — pale grey, pink, off-white. Quartz and feldspar first. Sierra-type granite, leucogranite, many rhyolites. Biotite as scattered books, not as a black felt.
- Mesocratic (35–65%) — true salt-and-pepper. Diorite, granodiorite, andesite. Hornblende and biotite share the face with plagioclase; the rock is not a granite from ten metres away.
- Melanocratic (65–90%) — dark grey to black, feldspar still findable with a lens. Gabbro, basalt, some pyroxene diorites. Already heavy in the hand.
- Ultramafic (CI > 90%; also called hypermelanic) — olivine, pyroxene, maybe a scrap of plagioclase. Peridotite, dunite, pyroxenite. Green-black, dense, no quartz argument to have.
Hololeucocratic (CI < 10%) is a finer split some authors use for alaskites and aplites. This page folds those into leucocratic, which is where they belong on a mapping traverse.
QAPF names and color index: two plots from one count
QAPF* (Streckeisen) names the rock from quartz, alkali feldspar, plagioclase and feldspathoid, recalculated to 100% after the mafics are thrown out. That is how granite, granodiorite, diorite, syenite, foidolite get their names. Color index never enters the double triangle.
Why bother with CI? Because two granites with the same Q–A–P point can be a pale muscovite leucogranite and a hornblende–biotite granite that would not be mapped as the same unit. QAPF answers “which felsic minerals?”. CI answers “how much dark stuff came along?”. They are orthogonal: the modal percentages feed QAPF, and those same percentages, sorted dark/light, feed CI. One thin section, two plots.
Porphyries: phenocryst colour is not the rock
A porphyry lies. Big white plagioclase phenocrysts on a dark andesitic groundmass make the hand specimen look paler than the magma that froze. Big pyroxene phenocrysts in a pale trachyte do the opposite.
The color index of the rock is still the whole-rock volume of dark minerals — phenocrysts plus groundmass, not whichever is prettier. Phenocryst mode first (area on a stained face, or a coarse grid), then the groundmass CI under the lens, then weighted. Reporting only the groundmass CI is honest only if it is said so; it is a different number, useful for the melt that remained, not for the name of the rock. Volcanic glass: a dark tachylitic mesostasis is mafic in appearance even when crystals cannot be counted. The glass is tagged dark or light by its actual colour, and “glass” is noted on the sheet.
Field estimate with a hand lens
Break a fresh face. Weathered skins go brown, green, white — anything but the mode. A ×10 lens on a medium-grained plutonic rock gets within about ±5–10% if the eye has been calibrated on counted specimens.
- Think in quarters at first (0, 25, 50, 75), then split the middle.
- Scan a palm-sized area, not the one crystal that caught the light.
- On fine-grained volcanics, trust the overall tone more than individual microlites that cannot be named.
- Carry one known granite (~10–20% CI) and one known gabbro (~70–85%) in the brain, not necessarily in the rucksack.
A field CI is a mapping letter, not a thesis table. Recount it on a stained slab or a thin section before a Streckeisen name is published.
How CI, density and SiO₂ usually walk together
Ferromagnesian minerals are heavier and poorer in silica. Felsic minerals are lighter and richer in silica. The three numbers usually walk together:
CI ↑ ⇔ ρgrain ↑ ⇔ SiO2 ↓
A Sierra or Cornish granite at CI ~12% and density ~2.67 g/cm³, SiO2 in the high sixties, is the textbook triple. A Columbia River or Icelandic basalt at CI ~70–85% (massive, not scoriaceous), ρ ~2.90–3.00, SiO2 ~50%, is the other end. Vesicular porosity breaks the density link without touching CI: the scoria is still melanocratic, it just has holes. Magnetite-rich gabbros break it the other way — CI already high, density higher still. When the triple disagrees, something besides “average granite versus average basalt” is going on. That is the useful disagreement.
Worked example: granite versus diorite
Two medium-grained plutonic rocks, no chemistry yet. Point counts on stained slabs (or the form on this page).
Rock A. Quartz 80, alkali feldspar 70, plagioclase 40, biotite 12. Total 202 points. Dark = biotite = 12. CI = 12/202 × 100 = 5.9% — leucocratic. Quartz is obvious; the rock is granite (sensu lato) before QAPF is even plotted.
Rock B — the default diorite on this form. Plagioclase 95, amphibole 55, biotite 20, quartz 15. Total 185. Dark = 55 + 20 = 75. CI = 75/185 × 100 = 40.5% — mesocratic. Quartz is accessory, hornblende is not. That is diorite territory, not granite, even before the An content of the plagioclase is asked.
CIA = 5.9% (leucocratic) · CIB = 40.5% (mesocratic)
The eye already knew. The count stops the argument on the outcrop. QAPF will refine granite versus granodiorite on A, and diorite versus quartz diorite on B; CI already split the two families.
Icelandic basalt on a BGS and USGS map
Massive Icelandic tholeiite is the field textbook of a melanocratic volcanic: CI commonly 70–85% when vesicles are gone, pyroxene + plagioclase, olivine in the more primitive flows, a hand-specimen density near 2.90–3.00 g/cm³. USGS flood-basalt units (Columbia River, Snake River Plain) occupy the same CI band. BGS maps of the British Palaeogene (Skye, Mull, Antrim) used “basalt” for the whole pile; the scoriaceous flow top is still melanocratic on CI and light in the rucksack because of vesicles.
A granite mapped next to that basalt — Cornish, Sierra, or a Skye epigranite — sits at CI 5–20%. The contrast is the mapping letter. IUGS wants the QAPF name as well; the color index is what the hammer saw first.
Accessory opaques, glass, and what counts as dark
Magnetite and ilmenite are dark, dense, and often only a few percent. Leave them out and a gabbro still looks like a gabbro; leave them out of an Fe-Ti ore gabbro and the CI becomes a joke. Glass: pale rhyolitic glass is light; brown to black basaltic glass is dark. A third bin called “mesostasis” is a problem unless a rule is written for it.
Apatite, zircon, titanite — accessories, usually ignored or lumped. Titanite is honey-brown; most people keep it off the mafic total unless it is rock-forming. The list of what went into “dark” is written once, at the top of the table, and the next thin-section session matches the last.
Weathered rinds lie: break a fresh face
Biotite goes gold, then white. Hornblende goes chlorite-green. Olivine goes iddingsite-red. Feldspar goes chalky. The rind of a melanocratic dyke can look mesocratic, and a granite saprolite can look like pale sand. Color index is a mode of the fresh rock.
CI is not estimated on a weathered joint face, a desert varnish, or a stream cobble with a brown skin. Hammer until the minerals have their proper colours, or wait for the thin section.
Alteration inside the rock is a different problem: a saussuritized gabbro is still a gabbro. Secondary phases are counted as dark or light by what they are, with a note if the primary CI was reconstructed from habits.
CIPW mafics versus modal color index
The CIPW norm rebuilds an anhydrous mineral list from the whole-rock chemistry: normative olivine, diopside, hypersthene, magnetite — no hornblende, no biotite, because CIPW will not spend water. Modal CI counts the amphibole and mica that actually grew.
A hydrous intermediate magma therefore shows more modal hornblende+biotite than CIPW pyroxene+olivine would lead one to expect, and the two “mafic percentages” will not match. That is not a lab error. CIPW is a chemical language; modal CI is a thin-section language. On volcanic glass, CIPW is sometimes all that is available. On a holocrystalline granite, the mode wins. Map them; do not average them.
Glossary
- Color index (CI)
- Volume percent of dark (ferromagnesian) minerals in the rock. IUGS cuts: leucocratic < 35%, mesocratic 35–65%, melanocratic 65–90%, ultramafic > 90%.
- Felsic
- Light-coloured, silica-rich minerals: quartz, feldspars, feldspathoids, muscovite. From feldspar + silica. Opposite of mafic on the color-index ledger.
- Mafic (ferromagnesian)
- Dark, Fe–Mg minerals: olivine, pyroxene, amphibole, biotite, and usually the opaque oxides. “Mafic” is also a chemical class (low SiO2); here it is the mineral list that goes into CI.
- QAPF
- The IUGS / Streckeisen double triangle that names igneous rocks from quartz, alkali feldspar, plagioclase and feldspathoid. Color index is a separate axis and does not move the QAPF point.