“…Three widely different models have been proposed suggesting that carbonatite magmas are: (1) a direct product of partial melting of mantle material (von Eckermann, 1948;Holmes, 1952;Dawson, 1962Dawson, , 1966Eggler, 1975;Koster Van Groos, 1975); (2) a differentiation product of a carbonated silicate liquid (King, 1949;King and Sutherland, 1960;Watkinson and Wyllie, 1971;Wyllie, 1987); (3) an immiscible carbonate liquid separated from an originally homogeneous carbonated silicate melt (von Eckermann, 1961;King, 1965;Koster Van Groos and Wyllie, 1966, 1968, 1973Middlemost, 1974;Cooper et al, 1975;Le Bas, 1977, 1981Freestone and Hamilton, 1980). The hypothesis of liquid immiscibility has recently been favoured by many authors, but a major objection is the contrasting evidence deduced from field and experimental relations.…”