Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)




Professor Denis M. Shaw


The Mamainse Point Formation formed on the margin of the Late Proterozoic Keweenawan Rift. It consists of olivine- and feldspar-phyric flood basalts with felsites and conglomeratic sediments, and spans the palaeomagnetic reversal normally regarded as the boundary between Lower and Middle Keweenawan. The basalts are classified as olivine and quartz tholeiites, although the dykes show some alkalic tendencies in trace-element characteristics and are better termed transitional basalts.

Formation of the lava pile was accompanied by low-grade burial metamorphism in the zeolite and prehnite-pumpellyite facies. Secondary minerals developed both within the rock and in veins and vesicles. A crude mineral zonation developed from epidote at the base to stilbite at the top of the exposed section. Heterogeneity, with the development of epidote-, chlorite- and albite-rich metadomains, is restricted to feldspar-phyric melaphyres in lower parts of the section. Calcite-laumontite veins are ubiquitous and even cross-cut the higher grade epidote-bearing rocks, suggesting that metamorphism took place in two stages, the second at a lower temperature. Geochemical evidence also suggests that fluid compositions changed. The initial hydration event established mineral zonation and resulted in addition of H₂O+ and alkalis (K, Rb, Li, Na, Ba) to the rocks. Cu was leached from some rocks and added to others. Latter fluids were more CO₂-rich depositing calcite and laumontite, with the resulting addition of CO₂ and mobilization of Ca and Sr.

Al, ΣFe, Mg, Ti, P, Y, Zr, Nb, Ni, Cr, Co, REE's, Th, Ta, Hf and U appear to be immobile or only moderately affected by the alteration. The stratigraphic variation of these elements allows the subdivision of the sequence into fiva series. The immobile trace-elements confirm the tholeiitic nature of the lavas and suggest an ocean floor to within-plate character. Incompatible element patterns suggest a similarity to E-type MORB.

Shallow-level fractionation of olivine-chromite-plagioclase can account for most of the geochemical variation in the five series, although in differing proportions. Clinopyroxene and Fe-Ti oxides crystallized after extrusion of the lavas. Variations in REE-patterns and incompatible element ratios cannot be explained by fractionation, however, and appear to have resulted during genesis of the primary magmas.

Modelling of REE patterns requires varying degrees of melting of a garnet-lherzolite mantle. However, two source compositions are indicated by initial REE concentrations, one with a relatively flat chondritic pattern and the other with a LREE-depleted pattern. Other incompatible element variations appear to be consistent with this, although Th and U abundances may point to further heterogeneities. The need for a different mantle source for virtually every series can be reconciled, however, by appealing to a dynamic partial melting model, in which variable batch and incremental melting take place within ascending mantle diapirs, each labelled with slightly different Th/U.

Petrographic and geochemical evidence suggest that some basalts have been contaminated by ?felsic crustal material, although this is restricted to series III flows and the dykes.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Included in

Geology Commons