Petrogenetic Models
Fractional crystallization – major elements
Mass balance: CPM = concentration in the parental melt C = concentration in differentiated melt CFCFCPMFMcum1 DM F = fraction of the melt remaining (1→0); CCii Lever rule: f PM FM ffc= (1– F) = degree of fractional fc ii crystallization CCcum FM
Vojtěch Janoušek:
Using whole-rock major- and trace-element data in interpretation of igneous rocks
(after Cox et al. 1979) Trends produced by fractional crystallization of cumulate (cum) consisting of one (P), two (P–Q) or three (P–Q–R) minerals. PM = primary melt, FM = fractionated magma
y Ax
Fractional crystallization – major elements Fractional crystallization – major elements
• Binary plots of major elements vs. SiO2 (Harker plots) or any other index of fractionation (e.g., MgO or mg#) often show linear relationships. Continuously evolving cumulate b Clinopyroxene C (9) • These trends on their own do not provide evidence for operation of fractional • Produces curved trends. S C (8) S (7) crystallization! C S C (8) wt.% L C (6) C (9) S L C (5) (5) C (7) • Partial melting or mixing would also produce linear trends (Wall et al. 1987). S C L C (3) L (4) L CaO C C (2) S L C (3) C (1) • However, the changes in fractionating assemblage result in inflections. S L C (2) C S 0 If present, they prove operation of fractional crystallization. C (1) y Ax S C (0) S Olivine
SiO2 wt.% Reverse modelling • For reverse modelling of fractional crystallization, the composition of the primitive magma is considered to be a mixture of differentiated magma and cumulus crystals. ii i CCfCPM cum fc FM(1 f fc )
(after Wilson 1989) • Numerical solution is provided, for instance, by the least-squares method (Bryan et al. 1969). Harker plots for a suite of cogenetic volcanic rocks developing by fractional crystallization of olivine, clinopyroxene, plagioclase and apatite.
1 Classification of trace elements Classification of trace elements
ionic potential = charge/radius
Determines, which elements will be mobile in hydrous fluids and thus strongly enriched in arc magmas (being contributed by the subducted plate) Non-conservative elements Large Ion Lithophile Elements (LILE): Cs, Rb, K, Li, Ba, Sr, Pb...
Conservative elements High Field Strength http://www.gly.uga.edu/railsback/FundamentalsIndex.html Elements (HFSE): Nb, Ta, Ti, Zr, Hf... (White 2005)
Presentation of trace-element data Spiderplots 0 • Binary plots Plotting Chondrite (Boynton 1984) Po- 1 (original) a b • Log–log plots • Arrange elements logically • Histograms (more incompatible on the left)
• Boxplots b • Divide each element’s R Rb (box and whiskers plots) concentration in the sample by (ppm) (ppm) • Box and percentile plots that in a reference material 200 400 600 800 100 • Plot y-axis using a log scale 0.05 0.10 0.20 0.50 ab 1 5 10 50 500 0.05 0.20 0.50 2.00 5.00 20.00 La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu 200 400 600 800 1000 1 5 10 50 500 Advantages/usage c Sr Sr • Elimination of the Oddo-Harkins Po- 1 (normalized) effect in the Solar System, the abundances of e
even-numbered elements are greater t i r d
than those of neighbouring odd- on h c E
numbered ones + abundances generally E R
/ e decrease with increasing atomic number. l amp • Spiderplots allow representing S much of the sample’s composition on a single graph. La Pr Pm Eu Tb Ho Tm Lu 1 10 100 Ce Nd Sm Gd Dy Er Yb
2 Spiderplots Spiderplots – examples
Standards for normalization 0 Standards for normalization MgO 5.8- 6 • Chondrites ( “Bulk Silicate Earth”) 6- 6.2 • Most primitive/least-altered sample 6.2- 6.4 6.4- 6.6 •Primitive Mantle 6.6- 6.8 6.8- 7 7- 7.2 • Mid-Ocean Ridge Basalts (NMORB) 7.2- 7.4 7.4- 7.6 7.6- 7.8 • Ocean-Island Basalts (OIB) Modifications 7.8- 8 8- 8.2 8.2- 8.4 • Averages of various crustal reservoirs, • Fields 8.4- 8.6 8.6- 8.8 8.8- 9 bulk, upper, lower… • Colour-coded by a certain 9- 9.2 a 9.2- 9.4 • Ocean Ridge Granites (ORG) 9.4- 9.6 parameter (SiO2, MgO…) La Pr Pm Eu Tb Ho Tm Lu 0.1 1 10 10 • Spider box plots Ce Nd Sm Gd Dy Er Yb • Spider box and percentile plots Rock/Chondrite (Boynton 1984) 0 0 1 0 1 b
Multielement plots for metavolcanic 0.1 1 10 rocks from the Devonian Vrbno La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Group, Silesia (Czech Republic) Chondrite-normalized (Boynton 1984)
(Janoušek et al. 2014) 1 REE patterns for dolerites from the Devonian Vrbno Group, Silesia La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Rock/NMORB (Sun and McDonough 1989) (SunRock/NMORB
Two types of trace elements Behaviour of accessories – solubility concept
Example – zircon “Dilute” trace elements (element partitioning) (Watson and Harrison 1983) • Trace elements that occur in small amounts in the crystal lattice, and their activity is proportional to their concentration (Henry 1803). • The coefficient of proportionality depends on the nature of the mineral but not Zr on the concentration of the element. C 12900 lnZrn 3.80 0.85( M 1) • Modelled by element partitioning between crystals and liquid. C Zr T L sat Essential Structural Constituents (solubility concept)
• Essential Structural Constituents (ESC) form a substantial part of the crystal T = temperature in K lattice of certain accessory minerals. Zr = 497644 ppm (49.7 wt. %) is the CZrn e.g., Zr in zircon, P in apatite or P, Th and LREE in monazite amount of Zr in an ideal zircon M = whole-rock chemical parameter • The Henry’s Law is not applicable. based on cation fractions: • Empirical approach, using experimentally determined mineral solubility in a Na K 2 Ca silicate melt. M Al Si
3 Partition coefficient Partition coefficient
HORNBLENDE CLINOPYROXENE GARNET 100 hbl/L cpx/L grt/L 50 KD KD KD rhyolite Partition coefficient compatible elements KD » 1 Depends on: concentrate into the mineral rather rhyolite dacite 10 • magma composition than in the melt basalt min/ L Cmin 5 dacite (acid/basic, water >1 : COMPATIBLE
K rhyolite D D K contents...) incompatible elements K « 1 1 dacite CL D basalt • temperature remain in the melt 0.5 basalt • pressure Partition coefficient 0.1 0.05 • mineral
< 1 : INCOMPATIBLE • mineral stoichiometry D 0.01 Does not depend on: La Ce Nd SmEuGd Dy Er YbLa Lu Ce Nd SmEuGd Dy Er YbLa Lu Ce Nd SmEuGd Dy Er Yb Lu (composition) K • concentration of the given element Measurement: plagioclase • oxygen fugacity concentration of any other elements 0.9 ab (Eu in plagioclase) • (Bulk analysis of mineral separates 10 in the system 5 and glass or matrix) 0.8 –13 fO2 =10 • In-situ measurement in –10 fO2 =10 0.7 1 –8 fO =10 phenocrysts and glass 2
(Sr) Bulk distribution coefficient
Pl/L 0.5D –6 D Pl/L
K fO =10 (ion probe, LA ICP MS,…) 0.6 2 Databases of KD values: • Experiments (doped material) ln K • Rollinson (1993) 0.1 DKd X 0.5 0.05 i i • GERM website • Calculations (lattice strain model) i (http://earthref.org/KDD/). 0.4 Pl/L Ln KD (Sr) = 1521/T – 9.909 0.01 6.8 6.9 7.0 7.1 7.2 La Ce Nd Sm Eu Gd Dy Er Yb Lu 104/T (K–1)
Partition coefficients Fractional crystallization
C0 = concentration in the parental melt Rayleigh equation C = concentration in differentiated melt C L L F (1)D F = fraction of the melt remaining (1→0); C (1– F) = degree of fc 0 D = bulk distribution coefficient for the crystallizing phases
‘Forbidden domain‘ Instantaneous solid: c 1 L cDcDcFinst (1) D c0 F sL0
Bulk cumulate: 1 F D ccs 0 Mineral/melt distribution coefficients for REE in dacites and rhyolites (Hanson 1978) 1 F
4 Fractional crystallization Equilibrium crystallization
The identification of fractionating phases is C = concentration in the parental melt Melt development: 0 facilitated by log–log plots of CL = concentration in differentiated melt whole-rock trace-element C F = fraction of the melt remaining (1→0); concentrations, in which the 0 CL (1– F) = degree of fractional originally exponential Rayleigh D FD(1 ) crystallization trends are converted to linear D = bulk distribution coefficient for the ones: crystallizing phases log(cL ) log(cD0 ) ( 1 ) log( F )
For granitoids are commonly Regardless of its exact mechanism used Rb, Sr, Ba whose (fractional/equilibrium), distribution coefficients (K ) are D crystallization quickly depletes relatively well known compatible elements from the melt. (Hanson 1978). Ba vs Sr patterns for the Kozárovice (diamonds) and Blatná (squares) intrusions of the Central Bohemian Plutonic Complex, Czech Republic (Janoušek et al. 2000) (White 2005)
Fractional crystallization – behaviour of Fractional melting accessories
C0 = concentration in the (unmelted) source Melt development: CL = concentration in the melt 1 1 F = degree of melting CS D (1F ) D = bulk distribution coefficient after C0 melting (residue)
At any moment, the instantaneous liquid equilibrates with the solid residue. The composition of a single melt increment is:
1 C 1 1 L (1F ) D CD0 Average composition of the (aggregated) melt:
C 1 1 L 1(1)F D CF0 Watson & Harrison (1984); Hoskin et al. (2000); Janoušek (2006)
5 Distinguishing between fractional Batch melting crystallization and partial melting
C0 = concentration in the (unmelted) source PM Melt development: 50 CL = concentration in the melt Co C F = degree of melting Fractional 0 F Pa rtial melti C r ng L D = bulk distribution coefficient after a 20
c DF(1 D ) t Batc i h
melting (residue) ) o
n β
a
l c 10 r y
s Ni ppm
t Log ( All the melt is formed, and then separated in a single batch. a l l i z Compatible a t 5 i o
n
FC ab2 100 200 500 Log (α) Sr ppm Incompatible Fractional crystallization (FC) produces an almost vertical trend whereas partial melting (PM) results in a nearly horizontal one (after Martin 1987).
Combined Assimilation and fractional (Binary) mixing crystallization (AFC)
Mass-balance equation: The AFC model assumes that the CM = concentration in the mixture extra heat needed for assimilation m Ck = concentration in the end-member k (which is an endothermic process) is Direct AFC models can be CfC () provided by the latent heat of M kk fk = proportion of the given end-member k calculated and plotted by k 1 crystallization. Petrograph (Petrelli et al. 2005) or special spreadsheets (Ersoy and Helvacı 2010; Trends in binary plots: Keskin 2013). • Element–element: straight line 1 vu 11, = concentration in • Element–ratio: hyperbola ba A C 11 theC assimilant L = concentration in • The sophisticated AFC the magma • Ratio–ratio: hyperbola [general] equations with many DC • Ratio–ratio: straight line [common = concentration in L parameters can be easily r = rate of assimilation to the crystallizing minerals denominator] fractional crystallization tweaked to yield solutions y=u/a C nicely reproducing the = initial concentration in 0 observed variation but vu the magma 22, geologically unrealistic ba 22 (Roberts and Clemens 2 1995) x=v/b
6 Wilson Partial melting – behaviour of accessories “Igneous petrogenesis” (1989)
Watson and Harrison (1984); Janoušek (2006)
concentration in source > melt concentration in source < melt
Inheritance source
is likely exhausted residue )
C0 ppm ( meltZr melt Zr Zr (ppm) Zr C C L sat L sat Inheritance is likely Inheritance to be re unlikely preserved sidue C0 a b 0 50 100 150 200 250 0 50 100 150 200 250
0.00.10.20.30.40.50.6 0.0 0.1 0.2 0.3 0.4 0.5 0.6
F F
Mid-Ocean Rifts Mid-Ocean Rifts
Ophiolite complexes • Provide sections of the former ocean floor
http://www.whitman.edu/geology/winter/JDW_PetClass.htm http://www.whitman.edu/geology/winter/JDW_PetClass.htm
7 Subduction zones Mid-Ocean Rifts (island arcs and continental arcs)
MORB = Mid-Ocean Ridge Basalt olivine tholeiite with low K2O (< 0.2 wt. %) and TiO2 (< 2.0 %)
N-MORB (normal MORB) shallow melting of a depleted mantle source:
Mg# > 65: K2O < 0.1 TiO2 < 1.0
E-MORB (enriched MORB) deeper, less depleted mantle source:
Mg# > 65: K2O > 0.1 TiO2 > 1.0
http://www.whitman.edu/geology/winter/JDW_PetClass.htm http://homepages.uni-tuebingen.de/wolfgang.siebel/lec/man.html
Anatomy of a subduction zone Geochemical behaviour of trace elements in (here continental arc) subduction zones
HFSE Subduction signal: B > As, Sb, Cs > Pb >Rb >Ba, Sr, Be ~ U ... Amph out LILE
Phl out
Winter (2001) http://www.gly.uga.edu/railsback/FundamentalsIndex.html
8 Continental vs. island arcs Anatomy of a subduction zone (compression)
Main differences of continental from island arcs: • Greater scope for crustal contamination while mantle- derived magmas ascend through Cotopaxi the thick, geochemically and isotopically evolved continental crust.
• Low density of the crust acts as a Moho (MASH) “density filter”, i.e. it decreases the ascent velocity (or even stops) the rising basic melts •Their stagnation leads to fractional crystallization and assimilation (MASH = Melting Assimilation Storage • Low solidus temperature facilitates Homogenization; Hildreth and partial melting of the fertile Moorbath 1988) at the mantle– continental crust lower crust boundary (at about (mantle wedge) the Moho depth)
Trace-element signature of subduction- Anatomy of a subduction zone (extension) related magmas
Subduction zone basalt
Moho (MASH)
(mantle wedge)
9 Trace-element signature of subduction- Ocean Island Basalts related magmas
(subducted slab dehydration esp. of serpentinite and basalt) (melting of subducted sediments)
Mantle component
Pearce et al. (2005) http://www.whitman.edu/geology/winter/JDW_PetClass.htm
Ocean Island Basalts Ocean Island Basalts
DMM Depleted Mantle DMM Depleted Mantle HIMU Subducted oceanic crust HIMU Subducted oceanic crust EM I modified subcontinental EM I modified subcontinental Treatise on Geochemistry Chap. lithospheric mantle lithospheric mantle 2.03: Sampling Mantle Heterogeneity through Oceanic EM II subducted sediments EM II subducted sediments Basalts: Isotopes and Trace Elements (A.W. Hofmann) FOZO “Focal zone” FOZO “Focal zone” (= deep mantle?) (= deep mantle?)
Hawaiian hot spot movement
http://www.geo.cornell.edu/geology/classes/Geo656/656home.html
10 Continental Rifts Continental collision
(Himalayas)
http://www-sst.unil.ch/research/plate_tecto/teaching_main.htm
Ocean closure Continental collision
Blueschist-facies
Coesite, Dora Maira, Alps: Photo C. Chopin
• Collision can be accompanied by deep subduction of continental crust into the upper mantle (even into stability fields of coesite and diamond) • This may lead to a strong contamination of the lithospheric mantle by geochemically evolved upper crustal material • Melting of such anomalous mantle domains Chopin ( 2003) may yield (ultra-) potassic magmas of http://geology.about.com peculiar composition
11 Post-Collisional Development Models for petrogenesis of granitoid rocks
Heat for crustal anatexis • Partial melting of crustal rocks • Stacking/thickening in-situ heat production by • Regional metamorphism – deep burial (granulite-facies) radioactive decay of K, Th, U • Injection of basic magma, basic magma underplating • Advection of heat by quickly • Crustal thickening or thinning exhumed hot lower crustal • Decompression meting during uplift of crustal rock complexes rocks or intruding basic magmas • Contamination of mantle-derived magmas • Conduction of heat from a assimilation of crustal material, followed by differentiation thermal anomaly in the • Differentiation of mantle-derived magmas mantle (slab break-off, mantle delamination, mainly by fractional crystallization asthenosphere upwelling, • Dehydration and partial melting of hydrated oceanic crust a mantle plume...), (including sediments) in subduction zones • Conduction of heat from anomalous mantle – in situ fluids and small-scale melts move upwards and trigger melting of the overlying radioactive decay of K, Th, U mantle wedge in crustally contaminated lithospheric mantle. • Partial melting of (meta-) basic rocks (amphibolites etc.) http://www.whitman.edu/geology/winter/JDW_PetClass.htm previous magma pulses, relicts of the subducted oceanic crust…
Petrogenetic classification of granitoid rocks Petrogenesis of granitic rocks
Alumina saturation and mineralogy Clarke (1992)
Peraluminous Metaluminous Peralkaline
Definition A > CNK CNK > A > NK A < NK
alumosilicates, Fe-rich olivine (fayalite), Characteristic cordierite, garnet, Pyroxene, aegirine, minerals topaz, tourmaline, amphibole, epidote arfedsonite, riebeckite spinel, corundum
Other common biotite, muscovite biotite, muscovite rare rare biotite minerals Pitcher (1983), Barbarin (1990), Pitcher (1993) Fe–Ti oxide phase Ilmenite Magnetite Magnetite http://www.whitman.edu/geology/winter/JDW_PetClass.htm
apatite, zircon, titanite, apatite, zircon, apatite, zircon, titanite, Accessories allanite, fluorite, cryolite, monazite allanite pyrochlore
87 86 ( Sr/ Sr)i 0.705–0.720 0.703–0.708 0.703–0.712
εNd << 0 ~ 0 variable
12 Geotectonic diagrams (basaltoids) Geotectonic diagrams (basaltoids)
IAT Island-arc Tholeiites CAB Calc-alkaline Basalts N-MORB N-type Mid-ocean Ridge Basalts E-MORB E-type Mid-ocean Ridge Basalts WVB WPT Within-plate Tholeiites EVB WPA Alkaline Within-plate Basalts
Wood (1980) Pearce (1982) Pearce (2008)
Geotectonic diagrams (granitoids) References and further reading
ALBARÈDE F. 1995. Introduction to the Geochemical Modeling. Cambridge University Press. ORG Ocean Ridge Granites BARBARIN, B., 1990. Granitoids: main petrogenetic classifications in relation to origin and tectonic setting. VAG Volcanic Arc Granites Geological Journal, 25, 227-238. WPG Within Plate Granites BOYNTON, W. V., 1984. Cosmochemistry of the rare earth elements: meteorite studies. In: Henderson, P. Syn-COLG Syn-Collision Granites (ed.): Rare Earth Element Geochemistry. Elsevier, Amsterdam, 63-114. BRYAN W.B., FINGER L.W. & CHAYES F. 1969. Estimating proportions in petrographic mixing equations by least-squares approximation.– Science 163: 926–927. CHOPIN, C., 2003. Ultrahigh-pressure metamorphism: tracing continental crust into the mantle. Earth and Planetary Science Letters, 212, 1-14. CLARKE, D.B. 1992. Granitoid Rocks. Chapman & Hall, London. COX K.G., BELL J.D. & PANKHURST R.J. 1979. The Interpretation of Igneous Rocks. George Allen & Unwin, London. DEPAOLO, D.J. 1981. Trace element and isotopic effects of combined wallrock assimilation and fractional crystallization. Earth Planet. Sci. Lett., 53, 189–202. ERSOY, Y. & HELVACI, C., 2010. FC-AFC-FCA and mixing modeler: a Microsoft© Excel® spreadsheet program for modeling geochemical differentiation of magma by crystal fractionation, crustal assimilation and mixing. Computers & Geosciences, 36, 383-390. HANSON, G. N., 1978. The application of trace elements to the petrogenesis of igneous rocks of granitic composition. Earth and Planetary Science Letters, 38, 26-43. HANSON G.N. 1980. Rare earth elements in petrogenetic studies of igneous systems. Ann. Rev. Earth Planet. Pearce et al. (1984) Sci. 8: 371–406.
13 References and further reading References and further reading
HILDRETH, W. & MOORBATH, S., 1988. Crustal contributions to arc magmatism in the Andes of PEARCE, J. A., HARRIS, N. B. W. & TINDLE, A. G., 1984. Trace element discrimination diagrams Central Chile. Contributions to Mineralogy and Petrology, 98, 455-489. for the tectonic interpretation of granitic rocks. Journal of Petrology, 25, 956-983. JANOUŠEK, V., 2006. Saturnin, R language script for application of accessory-mineral saturation models PETRELLI, M. et al. 2005. PetroGraph: a new software to visualize, model, and present geochemical in igneous geochemistry. Geologica Carpathica, 57, 131-142. data in igneous petrology. Geochemistry, Geophysics, Geosystems, 6, Q07011. JANOUŠEK, V. et al. 2000. Modelling diverse processes in the petrogenesis of a composite batholith: the PITCHER, W. S., 1983. Granite type and tectonic environment. In: HSÜ, K. J. (ed.): Mountain Central Bohemian Pluton, Central European Hercynides. Journal of Petrology, 41, 511-543. Building Processes. Academic Press, London, 19-40. JANOUŠEK, V. et al. 2014. Constraining genesis and geotectonic setting of metavolcanic complexes: a PITCHER, W.S. 1993. The Nature and Origin of Granite. Chapman & Hall, London. multidisciplinary study of the Devonian Vrbno Group (Hrubý Jeseník Mts., Czech Republic). ROBERTS, M. P. & CLEMENS, J. D., 1995. Feasibility of AFC models for the petrogenesis of calc- International Journal of Earth Sciences, 103, 455-483. alkaline magma series. Contributions to Mineralogy and Petrology, 121, 139-147. KESKIN, M., 2013. AFC-Modeler: a Microsoft® Excel© workbook program for modelling assimilation ROLLINSON H.R. 1993. Using Geochemical Data: Evaluation, Presentation, Interpretation. combined with fractional crystallization (AFC) process in magmatic systems by using equations of Longman, London. DePaolo (1981). Turkish Journal of Earth Sciences, 22, 304-319. WALL, V.J., CLEMENS, J.D. & CLARKE, D.B. 1987. Models for granitoid evolution and source compositions. The Journal of Geology, 95, 731–749. MARTIN, H., 1987. Petrogenesis of Archaean trondhjemites, tonalites, and granodiorites from eastern WHALEN J.B., CURRIE K.L., CHAPPELL B.W. (1987) A-type granites: geochemical Finland: major and trace element geochemistry. Journal of Petrology, 28, 921-953. characteristics, discrimination and petrogenesis. Cont. Mineral. Petrol., 95, 407–419. MCDONOUGH, W. F. & SUN, S., 1995. The composition of the Earth. Chem. Geology, 120, 223-253. WATSON, E. B. & HARRISON, T. M., 1983. Zircon saturation revisited: temperature and composition PEARCE, J. A., 1982. Trace element characteristics of lavas from destructive plate boundaries. In: effects in a variety of crustal magma types. Earth and Planetary Science Letters, 64, 295-304. THORPE, R. S. (ed.): Andesites; Orogenic Andesites and Related Rocks. John Wiley & Sons, WATSON, E. B. & HARRISON, T. M., 1984. Accessory minerals and the geochemical evolution of Chichester, 525-548. crustal magmatic systems: a summary and prospectus of experimental approaches. Physics of the PEARCE, J. A., 2008. Geochemical fingerprinting of oceanic basalts with applications to ophiolite Earth and Planetary Interiors, 35, 19-30. classification and the search for Archean oceanic crust. Lithos, 100, 14-48.
References and further reading References and further reading
WILSON, M., 1989. Igneous Petrogenesis. Unwin Hyman, London. WINTER, J. D., 2001. An Introduction to Igneous and Metamorphic Geology. Prentice Hall, Upper Saddle River, NJ. Janoušek V., Moyen J.-F., Martin H., Erban V., Farrow WOOD, D. A., 1980. The application of a Th–Hf–Ta diagram to problems of tectonomagmatic C. (in print): classification and to establishing the nature of crustal contamination of basaltic lavas of the Geochemical Modelling of Igneous Processes – British Tertiary volcanic province. Earth and Planetary Science Letters, 50, 11-30. Principles And Recipes in R Language. Springer Geochemistry 1. Springer, Berlin (July 2015) Web links • Geochemistry 455 (W.M. White, Cornell University) http://www.geo.cornell.edu/geology/classes/geo455/Geo455.html • Igneous and metamorphic geology (J. D. Winter, Whitman University) http://www.whitman.edu/geology/winter/JDW_PetClass.htm • Advanced petrology (J.-F. Moyen, University of Stellenbosch, South Africa) http://academic.sun.ac.za/earthSci/honours/modules/igneous_petrology.htm • EarthRef.org. The website for Earth Science reference data and models http://earthref.org/
14