The indigenous gossans of Minas Carlota formed in situ by the oxidation of steeply-dipping massive, banded pyritic ore bodies from 8 to 20 ft. thick and 100 to 1200 ft. long. The gossans are apparently younger than the dissected summit peneplain of the Trinidad Mountains and probably younger than the laterites of eastern Cuba. They are pseudomorphic with a banding inherited from the primary ore and are divisible into 3 gradational types by means of color, weight, and porosity. An "earthy" layer separates gossan from sulfide. Locally, ground water has removed the earthy layer and formed caverns. The assumption that earthy layers are transitional products and that most gossans have collapsed has led to a neglect of relict structures in gossans. At Carlota the earthy layer is not transitional and the relict structures show that the ore bodies are commonly en echelon. Specific gravities of limonite were determined to ascertain whether the conversion pyrite-->limonite was accompanied by expansion or contraction ("collapse"). Results were not consistent. Discrepancies are attributed to the presence of unoxidized pyrite, entrapped air, hematite, or extraneous material. In general a direct relationship exists between gossan porosity and the percentage of original sulfide. Exotic gossan (limonite) may coat indigenous gossan or occur as float. Varieties include: breccia, boxwork, stalactitic drip, and spongy cinder. Indigenous gossan, except where recently broken, does not occur as float. Float gossan neither indicates the type of indigenous gossan present nor the type or grade of ore to be expected. Chemical analyses indicate a general relationship between a high Fe and a low S content, a possible relationship between high Fe and low P, and a general relationship between high S and low P. The Se and precious metal content could be used to calculate the erosion the region has undergone. Differential thermal and X-ray analyses show that goethite, with or without hematite, is the dominant iron oxide. The hematite may have formed by dehydration of limonite at the surface or by reaction between bacterially-produced ferric nitrate and ammonia. Absence of siderite bears out the Eh-pH calculations of Kelly and others. Related problems are reviewed and some hypothetical gossans outlined.