Petrology and lithogeochemistry of the Wildcat Brook Mo-W deposit, Charlotte County, New Brunswick, Canada: insights into log-ratio analysis.

The Wildcat Brook Mo-W deposit is located approximately 9 km east of the former Mt. Pleasant Sn-W-Mo mine in Charlotte County, New Brunswick. It is hosted by a leucocratic, quartz-feldspar porphyritic-to-aplitic, peraluminous, EW-striking, moderately north-dipping dyke intruding turbiditic metasedimentary wackes/argillites of the Fredericton Trough north of the Saint George batholith. High-grade Mo mineralization to 4 % (per metre) occurs in two niches: (i) molybdenite blebs to 4 mm disseminated in miarolitic cavities within albite- and muscovite-altered dyke, and (ii) medium-grade Mo mineralization at the margins/cores of 2-5 cm wide quartz veins cutting altered/ unaltered dyke and adjacent metasedimentary rocks. The mineralized dyke has been intersected by 17 diamond drill cores spanning approximately 250 m along both strike and dip. Dyke thickness ranges from 17 to 42 metres (27 m average), with a length-weighted average grade of 0.27 % Mo. Two styles of hydrothermal alteration occur in the dyke: albite and muscovite-quartz. Molar element ratio (MER) analysis of 110 drill core samples, constrained by petrography, reveal the addition of Na and loss of K and Ca during albite alteration: Microcline + Na+ = >Albite + K+, Anorthite + 4 Quartz + 2 Na+ = >2 Albite + 3 Ca+2, and the addition of K and loss of Na during muscovite plus quartz alteration: 3 Albite + K+ + 2 H+ = >Muscovite + 6 Quartz + 3 Na+. Because hydrothermal alteration recognized in the dyke can be described by balanced chemical reactions, these processes are mathematically linear. Furthermore, when plotted on MER diagrams the data continue to form linear trends. Unfortunately, when plotted in logarithmic space, as log ratios, patterns become decidedly non-linear, significantly complicating data interpretation. Additionally, the use of linear statistical procedures, such as principal components or regression analysis, on such log-transformed data is numerically invalid, and so is not an appropriate way to understand the causes of compositional variations in rocks. [ABSTRACT FROM AUTHOR]