ABSTRACT Massive, boudinaged, rose-coloured layers up to one metre in thickness composed of 60–80 vol% spessartine-rich garnet and 15–40 vol% quartz occur in high-pressure, garnet-mica schist on Margarita Island, Venezuela. In the past, such spessartine-garnet-rich rocks have often, although inaccurately, been called ‘coticule’. Apatite can be significant, reaching 20 vol%. Accessory minerals are allanite, rutile (commonly rimmed by ilmenite) and rare phengite. Within the same mica-schist sequence, intercalated calc-silicate layers from one to two metres in thickness exhibit alternating layers, on a sub-mm- to cm-scale, of quartz and greenish layers rich in zoisite, epidote-supergroup minerals and diopside. Hyalophane and barite are present in small amounts. Calc-silicates are commonly pitted, with yellow to dark-brown, gossanous, botryoidal crusts. The composition of garnet in the garnet-rich layers, the calc-silicate layers and their host metasediments is distinctly different. In the garnet-rich layers, compositions are almandine-dominant (Alm57-63), followed by spessartine (Sps18-24), grossular (Grs2-16), pyrope (Prp6-11) and andradite (Adr0-6). Garnet in calc-silicate layers can range from almost pure grossular to Grs45-49Alm32-34Sps13-15, and in the enclosing metasediments is typically Alm71-73Prp14-15Grs3-8Sps4-7And2. Chondrite-normalized REE patterns of all three rock types are similar in shape, differing only in absolute contents, so that a common sedimentary precursor is indicated. Major and trace element chemistry of the garnet-rich layers are best explained if their pre-metamorphic protolith originated as a mixture of 60–70% normal detrital input and a precipitate from a distal source of hydrothermal activity. The term meta-exhalite is appropriate. The calc-silicate layers are more enigmatic, but an origin as a mixture of a hydrothermal component with a Ca-rich sediment also appears likely. Numerous lines of evidence indicate that the protoliths of the high-pressure rocks of Margarita Island originated in an Early Cretaceous back-arc basin in NW Colombia, so that an exhalite origin for garnet-rich and calc-silicate rocks is a plausible scenario.