Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A comparative study of lherzolite nodules in basaltic rocks from British Columbia Littlejohn, Alastair Lewis 1972-12-31

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


UBC_1972_A6_7 L58.pdf [ 8.19MB ]
JSON: 1.0053109.json
JSON-LD: 1.0053109+ld.json
RDF/XML (Pretty): 1.0053109.xml
RDF/JSON: 1.0053109+rdf.json
Turtle: 1.0053109+rdf-turtle.txt
N-Triples: 1.0053109+rdf-ntriples.txt
Original Record: 1.0053109 +original-record.json
Full Text

Full Text

A COMPARATIVE STUDY OF LHERZOLITE NODULES IN BASALTIC ROCKS PROM BRITISH COLUMBIA by ALASTAIR LEWIS LITTLEJOHN B.Sc, University of Aberdeen, 1969 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE In the Department of Geology if We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 1972 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. ALASTAIR L. LITTLEJOHN Department of GEOLOGY The University of British Columbia Vancouver 8, Canada Date <2)9th. Februry 1972 .ii ABSTRACT Lherzolite nodules in basaltic rocks from three localities in British Columbia include rocks of mantle origin and crystal cumulates. Partial chemical analyses show that the compositional ranges of the minerals are narrow for both major and minor elem ents and fall within the ranges reported for lherzolite nodules elsewhere. Each suite is characterised by a definite range of concentrations of some elements. Olivine in nodules from Castle Rock and Jacques Lake show fabrics resulting from deformation in the solid state prior to their incorporation into their host rocks but those from Nicola Lake are undeformed. The distribution of iron and magnesium between coexisting phases is examined using an ideal ionic solution model. Differences in the distribution coefficients between the suites are probably due to different temperature and pressure conditions at the source of the nodules. The distribution of iron and magnesium between coexisting spinel and olivine gives nominal temperatures of formation of 838°C for Nicola Lake nodules, 1085°C for Jacques Lake nodules and>l600°C for Castle Rock nodules. Differences among the suites in the distribution, of Ni, Co, Mn and Zn between coexisting silicates are independent of variations in composition and are apparently due to different conditions of formation. The Castle Rock and Jacques Lake lherzolites are residual fragments of the upper mantle left after extraction of an under-saturated basaltic liquid from parental mantle rock. The source of the Castle Rock nodules probably lies at greater depth than that of the Jacques Lake nodules. The Nicola Lake nodules are crystal cumulates and formed at an early stage of basalt genesis within the upper mantle or lower crust. TABLE OF CONTENTS .ill CHAPTER 1 CHAPTER 2 CHAPTER 3 CHAPTER 4 CHAPTER 5 CHAPTER 6 CHAPTER 7 CHAPTER 8 CHAPTER 9 CHAPTER 10 BIBLIOGRAPHY APPENDIX 1 APPENDIX 2 Page Introduction. 1 The Localities of Nodules in British Columbia. 4 (a) Nodule localities. h (b) Petrography of the host rocks. 4 Petrography of the Nodules. 9 (a) General. 9 (b) Petrographic descriptions. 9 (c) Secondary textural features. lh Petrofabric Study. 20 (a) General.(b) Description of the fabrics. 21 (c) Discussion. 26 Mineral Compositions. 3(a) General. 36 (b) Olivine. 37 (c) Orthopyroxene. 3(d) Clinopyroxene. 4o (e) Spinel. The Distribution of Iron and Magnesium Between Coexisting Minerals. 4-8 (a) Theory. 4(b) Results for coexisting silicates. 50 (c) The effects of temperature and pressure. 52 (d) The distribution between spinel and olivine. 6 (e) Results for coexisting spinel and pyroxenes. 6l The Distribution of Trace Elements Between Coexisting Silicates. 63 (a) -Theory. 6(b) The distribution of Ni.Mn.Co and Zn. 65 (c) Other elements. 73 The Origin of the Nodules. 76 (a) Temperature and pressure. 7(b) The nature of the source. 83 The Upper Mantle in British Columbia. 9^ Conclusions. 98 100 Analytical Techniques. 107 Error Propagation in Temperature Calculations. 112 LIST OP TABLES .iv Table. Page 1 Modes of Nodules with Analysed Minerals. 36 2 Partial Chemical Analyses of Olivines. 38 3 Partial Chemical Analyses of Orthopyroxenes. 39 4 Partial Chemical Analyses of Clinopyroxenes. 4l 5 Partial Chemical Analyses of Spinels. 45 6 Values of K-q for Coexisting Olivine and Pyroxenes with Analysis of Variance. 51 7 Values of Kp for Coexisting Spinel and Olivine with Analysis of Variance. 58 8 Temperatures of Formation of Coexisting Spinel and Olivine. 59 Values of K„ for Coexisting Spinel and Pyroxenes with Analysis of Variance. 62 10 (Tr/Cr) Ratios of Analysed Minerals. 67 11 Trace Element Distribution Coefficients. 70 12 Analysis of Trace Element Variance between Jacques Lake and Castle Rock Suites. 71 13 Values of k™ and kc for Coexisting Pyroxenes with Analysis of Variance. 75 LIST OF FIGURES Figure - Page 1 Localities of Ultramaflc Nodules In British Columbia. 5 2 Basalt Coating around Lherzolite Nodule. 7 3 Modes of Ultramaflc Nodules from British Columbia. 10 4 Aggregate of Spinel and Clinopyroxene. 12 5 Exsolved Clinopyroxene in Orthopyroxene. 12 6 Exsolved Spinel in Orthopyroxene. 13 7 Exsolved Spinel in Olivine. 18 Mineralogical Banding in Castle Rock Lherzolite. 15 9 Alignment of Sheared Spinel Grains. 15 10 Reaction of Orthopyroxene at Margin of Host. 17 11 Porous-looking Outer Rim of Clinopyroxene. 17 12 Fluid Inclusions in Olivine. 19 13 Olivine Fabric Diagram for JL-A. . .23 14 Olivine Fabric Diagram for JL-50. 23 15 Olivine Fabric Diagram for JL-24. 24 16 Strained Olivine with Abundant Kink Bands in Jacques Lake Lherzolite. 25 17 Olivine Fabric Diagram for NL-8. 24 18 Mutually Interfering Grain Boundaries between Olivines and Orthopyroxene in Nicola Lake Lherzolite. 25 19 Olivine Fabric Diagram for S-2. 28 20 Olivine Fabric Diagram for S-4. , 28 21 Olivine Fabric Diagram for CRi8. 29 22 Strain-free, Recrystallised Olivine in Castle Rock Lherzolite. 30 LIST OP FIGURES (continued) vi Figure Page 23 Enstatlte Fabric Diagram for CR-8. 29 24 Variation of Na20 and Ti02 with Al20o in. Clinopyroxene. 42 25 Variation of Al20o ,CaO and Na20 with MgO:FeO in Clinopyroxene. 44 26 Composition of Analysed Spinels. 4? 2? Distribution of Iron and Magnesium between Coexisting Olivine and Orthopyroxene. 54 28 Relative Proportions of Zn, Co, Nl and Mn between Coexisting Olivine, Orthopyroxene and Clinopyroxene.68 29 Relative Proportions of Pb and Cu between Coexisting Olivine, Orthopyroxene and Clinopyroxene.74 30 Composition of Analysed Pyroxenes in terms of MgSi03 - CaSi.03 - AI2O3. 78 31 Variation of AlgO^ between Spinel and Pyroxenes. 80 32 Relative Stabilities of Various Ultramafic Mineral Assemblages. 81 33 Various Lherzolite Solidi. 90 34 Part of the Liquidus Diagram of the System Fo - Di - Si02 at 20Kb. Pressure. 93 35 Major Structural Features Related to Recent Volcanism in British Columbia. 7 .vii ACKNOWLEDGEMENTS I would like to express my thanks and appreciation to Dr. H.J. Greenwood for suggesting this topic and supervising the work. Thanks are also due to Dr. J. Souther who provided the Castle Rock specimens, to Dr. K. Fletcher who gave his advice freely on atomic absorption techniques and to Mr. J. Harakal who did most of the probe work. The advice and help of Dr. R. Delavault, Miss S. Barr and Mr. A. Dhillon during the analytical portion of the study, and that of Mr. C, Fletcher on the use of computers is greatly appreciated. The loan of a typewriter from Mr. G. Carglll was invaluable. The receipt of an N.R.C. Post-graduate Scholar ship during the period of this study is greatly appreciated. CHAPTER 1 Introduction .1 Ultramaflc nodules occur in basaltic rocks throughout the world (Forbes and Kuno 19651 196?). The most common type of nodule is spinel peridotite, consisting of various proportions of olivine, diopside,erisjatlte and spinel. Garnet, hornblende, phlogopite, anorthite and more rarely melilite and leucite also occur as primary phases in ultramaflc nodules (Green 1968). Gabbroic and granulitic rocks derived from the subvolcanlc basement often occur in association with the ultramaflc types. Gabbro appears to be the commonest type of nodule found in basalts of all types, but ultramaflc nodules usually occur in basaltic rocks of alkalic affinities- (White 1966; Forbes and Kuno 1967). Only rarely are they found in tholeiitic basalts. Despite their relative scarcity, ultramaflc nodules are important since they may yield information on the nature of the upper mantle and on the genesis of basalts. Ross et. al. (1954) first drew attention to the relationship between ultramaflc nodules, dunites and the upper mantle. They found that ultramaflc. nodules - jaround^ the world have a uniform mineralogy and'chemistry and suggested that the nodules were derived from a uniform mantle peridotite. However, White (1966), Jackson (1968) and Kuno (1969) studied suites of nodules from Hawaii and found that they could be grouped according to mineralogy, chemistry and the character of the host rock. These authors have divided Hawaiian nodules into four groups. These are (a) a lherzolite series which forms part of the .2 upper mantle; (b) a dunite-wehrlite-pyroxenite series which forms crystal cumulates in the lower parts of the Hawaiian magma reservoirs; (c) an eclogite series which forms pockets in the dominantly lherzolitic mantle; (d) a gabbro series which forms part of the crust. Studies by Yamaguchi (1964), Kuno (1967), Aoki (1968) and Ishlbashi (1970) have shown that there is also a wide variety of nodule types in Japan. No single hypothesis seems adequate to explain the origin of all ultramafic nodules. Several hypotheses have been advanced to explain the origin of these rocks, tnSLuding^(a) they are fragments of the mantle; (b) they are crystalline residues of partially melted mantle; (c) they are products of crystal settling, formed during ascent of their basaltic hosts; (d) they are products of crystal settling, formed at some early stage of basalt formation, not necessarily the present host rock; (e) they are.fragments of some earlier formed ultramafic body in the crust. That any one' of the above hypotheses holds for all nodules is not generally accepted. However there is still controversy over the origin of any particular type of nodule. Lherzolite, the commonest ultramafic variety, is of particular interest as it is analogous to "pyrolite", a hypothetical mantle rock (Ringwood 1966, 1969). White (1966) considers lherzolite nodules to be residue from fusion of the primitive mantle. Jackson (1968) considers lherzolite to be part of a heterogeneous mantle and dunlte to be the refractory residue. Carter (1970) suggests that most lherzolite nodules are refractory residue from the upper mantle but that those relatively rich in iron .3 are cumulates. Kuno and Aoki (1970) have found a wide variety of lherzolite compositions throughout the world and suggest that this is a result of different degrees of partial melting in the mantle which is itself composed of lherzolite with a relatively low Mg«Fe ratio. O'Hara (1963, 1967. 1968) and Brothers (i960) support the hypothesis that lherzolite and other nodules are cognate. An accidental origin is considered, in general, to be unlikely, although Instances are known where ultramaflc nodules, including lherzolltes, are thought to be xenollths of sub-volcanic stratiform complexes (Fuster et. al. 1969). In order to examine some of the problems outlined, particularly with regard to lherzolite, suites of nodules from British Columbia have been collected and studied by the writer. The objectives were twofoldj (a) to describe the occurance and types of nodules which have been found in British Columbia; and (b) to determine the differences and similarities between suites and relate these if possible to the source of the nodules. The nodules and host rocks were studied by standard petrographic techniques. Petrofabric studies on the olivine and enstatite of the nodules were carried out using a 4-axes universal stage. The minerals of the nodules were separated and analysed by means of atomic-absorption spectrophotometry, electron microprobe and by wet-chemical means. CHAPTER 2 The Localities of Nodules In British Columbia. (a) Nodule localities. The localities of ultramaflc nodules which have been documented in the literature in British Columbia are shown on Pig. 1. Of these, suites from Castle Rock, Jacques Lake and Nicola Lake were studied. The host rocks were not studied in detail. (b) Petrography of the host rocks. (1) Castle Rock. Castle Rock is is a small peak on the northern flank of the Klastllne Plateau, kO miles east of Telegraph Creek in Northern British Columbia. It is one of several Quaternary volcanic centers which occur in the region (Souther 197°)• Rounded nodules from*2 to 6 inches in diameter are found in a volcanic breccia, made up of sub-rounded fragments of a black fine-grained alkali basalt in a matrix of palagonite. The contact of the nodules and breccia is sharp; in some specimens there Is a thin film of basalt coating the nodule. This film, petrographically identical to the basalt fragments, consists of small laths of plagioclase (An£0) set in a matrix of glass.olivine, magnetite and minor clinopyroxene. The film appears to have prevented the ." ; S:w-disintegration of the nodules during the explosive extrusion of the breccia. The breccia also contains small xenollths of diorite; presumably these are fragments of an underlying intrusive body. (ii) Jacques Lake. The nodules at the Jacques Lake locality are found in a 1 Castle Rock 2 Jacques Lake 3 Haggen's Point 4 Boss Mountain 5 Nicola Lake 0 (00 200 300 miles small dissected cone 4 miles south of Quesnel Lake in Central British Columbia. The host rock is a coarse tuff exhibiting crude layering in places. The tuff is made up of rock fragments of various types and sizes, cemented by a brownish-green matrix consisting of small rock fragments and partly devitrifled glass. The rock fragments consist of sedimentary, plutonic, metamorphic and volcanic xenoliths which are.presumably representative of the crust beneath Jacques Lake. No single type is predominant and the size and frequency of each kind are highly variable. Ultramafic nodules, mainly lherzolite, are found infrequently throughout:*: the cone. -Many of the-nodules are coated with a thin film of basalt consisting of small laths of altered plagioclase in a fine grained matrix of glass.magnetite and olivine (Fig. 2). This coating appears to be the original basalt within which the nodules were suspended before the extrusion of the tuff and has served a similar purpose to the coating around the Castle Rock nodules. The nodules are generally rounded or sub-rounded and range from 1 to 15 inches in diameter. Most are less than 6 1" inches in diameter. The cone is Quaternary in age and appears to be similar'to several other cones and flows of alkali basalt which occur in the area (Cambell 1961). (iii) Nicola Lake. • Ultramafic nodules are found in scattered boulders 3 miles south of Nicola Lake in Central British Columbia. The host rock is a, dark grey, fine-grained, vesicular basalt. It consists Fig. 2 Basalt coating around lherzolite nodule in Jacques Lake tuff. .8 of small laths of unzoned plagloclase (AngQ.), rounded olivine grains , interstitial glass and minor magnetite. Often partly-corroded xenocrysts of olivine and pyroxene are found. The basalt is alkaline in character. The age of the basalt is unknown but is possibly Tertiary. The boulders appear to have been brought to their present position by glacial action from the north where there is a large volume of Tertiary basalt. (K. C. McTaggart pers. comm.). .9 CHAPTER 3 Petrography of the nodules, (a) General, All the nodules studied consist of various proportions of olivine, clinopyroxene, ..orthopyroxene and spinel. Olivine is the dominant phase; orthopyroxene generally exceeds clinopyroxene; spinel is a minor phase. The modes of 29 nodules from the three localities are shown on a tertiary diagram (Fig. 3)» F°r this purpose spinel is omitted but all nodules contain from 0.5 to 2% (by volume) spinel. Modes were calculated by point-counting from 30 to 300 grains in thin section, depending on the size of the nodule. Some of the results for the smaller nodules may not be accurate (particularly those from Nicola Lake),but are included for comparison. Modes of other nodules from British Columbia are shown for comparison. The data are taken from Soregaroli (196?) and Tredger (1969) for Boss Mountain and Haggen»s Point > respectively. . All the nodules studied are lherzolites. The nomenclature is in accordance with the classification of Jackson (1968). According to White (1966) and Kuno and Aoki (1970) lherzolite is the predominant ultramafic rock found as nodules at other localities around the world. (b) Petrography of the nodules. All the nodules from each suite have allotrlomorphic granular textures, although there are some differences between the suites. The Castle Rock and Nicola Lake nodules are medium-Dunite Orthopyroxenite Fig. 3 Modal composition of ultramafic nodules from British Columbia. • Jacques Lake O Boss Mountain a Castle Rock © Haggen's Point S Nicola Lake .11 grained, whereas the Jacques Lake ones are coarse-grained and are very friable. Olivine forms an interlocking mosaic of rounded grains ranging in size from 0.5 to 2.00 mm. in diameter. In many of the Nicola Lake specimens some grains have mutually Interfering boundaries. In some of the Jacques Lake specimens olivine may be up to 5*00 mm. in diameter. In all suites it is pale green. Orthopyroxene forms subhedral grains some of which are larger -than the olivine and partly enclose it and others that are small, anhedral and are interstitial to the olivine. It-is dark brown in hand specimen and colourless in thin section. Clinopyroxene forms small anhedral grains, generally hot more.than 0.5 mm. in diameter, interstitial to both olivine and orthopyroxene. It tends to occur as aggregates of three or four grains. It is a bright emerald green in hand specimen and colourless or pale green in thin section. Spinel occurs as irregularly shaped grains interstitial to and partly enclosing the silicates, particularly the pyroxenes (Fig. 4). It is black in hand specimen and reddish-brown in thin section. Exsolution lamellae are found in some of the silicates. Lamellae of clinopyroxene In orthopyroxene (parallel to ^100} of the host) are found only In the largest orthopyroxenes. They are thin and pinch out towards the boundary of the host (Fig. 5). Both pyroxenes and very rarely olivine contain thin lamellae of a brown isotropic mineral which is presumed to be spinel (Figs. 6, 7). In the pyroxenes the lamellae are sub-paraaiel Fig. 4 Aggregate of spinel (brown) and clinopyroxene (grains with cleavage); grains with no cleavage showing are olivine; specimen S-2. X-30; plane polarised light. Pig.5 Exsolution of clinopyroxene in orthopyroxene (dark); specimen JL-50* X-3O; crossed nicols. .13 Fig. 6 Exsolved spinel (brown) in orthopyroxene (grains with cleavage); specimen S-3. X-30; plane polaeised light. Fig. 7 Spinel (brown lamellae) exsolved in olivine (yellow); specimen JL-50. X-30; crossed nicols. .14 to {001^ and In the olivine they are parallel to the (010) cleavage. Many of the nodules show signs of deformation such as kink bands in the olivines. A discussion of deformation follows in Chapter 4. Some specimens from Castle Rock are layered. Specimen s-4 has two well-defined bands, each about 5 nun. thick, of clinopyroxene which are separated by a thicker band of olivine (Fig. 8). In other nodules there is an ill-defined layering marked by thin disseminated stringers of spinel (Fig. 9) or orthopyroxene. Other nodules are massive. Specimens from Jacques Lake do not in general show layering. Rarely, banding similar to that observed in S-4 above was seen in the field. Unfortunately these specimens could not be broken out of the host tuff. None of the Nicola Lake nodules show layering but their small size might make this difficult to see. (c) Secondary textural features. Most of the nodules from all three suites show evidence of disequilibrium between the enclosing rock and the primary minerals. Usually reaction has taken place at the margin of the nodule', although in some specimens (particularly from Jacques Lake) basalt has been able to permeate the whole rock so that the minerals of the interior have been affected. Olivine shows the least effects of reaction with the host. Where In contact with basalt, the margin of the olivine grains commonly shows a fine-grained rim of secondary olivine and magnetite. .15 Fig. 8 Mineralogical banding in specimen S-4 from Castle Rock. The bright green bands are diopside; lighter green is olivine. * ** i " . -J. - t ' j 7. Fig. 9 Sheared spinel (brown) cutting through silicates; specimen S-2 from Castle Rock. .16 Spinel grains in contact with the host have a dark margin, presumed to magnetite. Orthopyroxene, where in contact with the host rock, has a rim of rim of finely divided material of high birefringence set in a dark cryptocrystalline matrix (Fig.10). This appears to be a result of lncongruent melting of orthopyroxene which has produced olivine and glass. The clinopyroxene of some nodules in contact with the host rock has a thin rim of secondary clinopyroxene. In the Jacques Lake nodules and:; rarely in those from Castle Rock, the clino-pyroxenes have a porous-looking outer zone which may be up to a third of the diameter of the grain in width (Fig. 11). The entire grain extinguishes uniformly and has uniform birefringence. The spongy zone is riddled with an extremely fine-grained dark material which appears to be partially devitrifled glass. This rim is entirely different from the rims observed at the margins of the nodules which are of secondary pyroxene with different extinction from the parent. According to White (1966) similar features are a result of a depletion in the jadeite component of the pyroxene thuss jadeite-bearing diopside —> jadeite-poor dlopside + feldspar The reaction rims occur throughout the rock and for reasons given in Chapter 8, they are considered here to be a result of partial melting. All the reaction phenomena described here have been reported elsewhere (Wilshlre and Binns 1961j Talbot et. al. 1963; .17 Fig. 10 Reaction of orthopyroxene at margin of host. The highly birefringent material is secondary olivine; specimen S-4. X-100; crossed nicols. Fig.11 Porous-looking outer rim of clinopyroxene (on right). The orthopyroxene (on left) is unaffected. The mineral with no cleavage showing is olivine; specimen JL-24. X-30; plane polarised light. .18 Yamaguchi 1964; White 1966; Kutolin and Prolova 1970). The reaction textures indicate disequilibrium between nodules and magma at a high crustal level but do not preclude equilibrium at greater depth. Fluid inclusions were found in olivines in most nodules from all the suites. Irf some cases they were also observed in pyroxenes. Many of these are two phase (gas + liquid) inclusions. Roedder (1965) has found two phase fluid inclusions in olivines from many localities. Most of these are C02; H20 is found in some nodules. The composition of the inclusions in the B.C. nodules is unknown. Most of the inclusions are found along fractures (Fig. 12) and along the cleavage of the olivine. In some cases planes of inclusions cut across fractures (Fig. 12). They are most abundant where the basalt has penetrated the nodule. They thus appear to be secondary and may have formed after the nodules were captured by their hosts. Some Inclusions may be primary, \ but distinguishing these from secondary ones is uncertain. Fig. 12 Fluid Inclusions aligned along fractures in olivine; specimen JL-50. X-400. .20 CHAPTER 4 Petrofabrlc Study. (a) General. Previous work on ultramafic nodules has revealed that the olivine in many nodules has a prefered orientation (Turner 1942} Brothers 1959,I960j Talbot et. al. 1963s Collee 1963; Black and Brothers 1965; Brothers and Rodgers 1969). These authors have shown that there is a variety of olivine fabrics in nodules and have attempted to draw analogies between the olivine orientation patterns found in nodules and those found in ollvine-bearing rocks from other environments. Thus Brothers (I960) compared olivine fabrics in nodules to those in flow-banded troctolites and basic dykes and suggested the nodules formed by crystal settling in a moving magma. Other workers consider that the fabrics of olivine from nodules are similar to those of rocks which have been deformed. In order to determine whether olivines from nodules in British Columbia have a prefered orientation, and if so, whether such an orientation could be related to the source of the nodules, fabric diagrams for olivine of three nodules from Castle Rock, three from Jacques Lake and one from Nicola Lake were prepared. An enstatite fabric diagram of one of the Castle Rock nodules was also prepared. The orientation of the olivine and enstatite principal optic directions was measured on a 4-axes universal stage. Two of these were determined in this way, the third being found by construction. The diagrams were constructed on the lower .21 hemisphere of a Schmidt net. Measurements were made on every grain intersected on suitably spaced lines of traverse in order to minimise sampling errors, (b) Description of the fabrics. Figs. 13, 14 and 15 are the fabric diagrams for the Jacques Lake nodules. All three nodules have similar olivine fabrics, although there are some differences between each one. The main element of the fabrics is the three mutually perpendicular maxima. In specimen JL-24 (Fig. 15) and JL-A (Fig. 13) this is modified. In these fabrics (3 forms a weak partial girdle. In addition JL-24 (Fig. 15) has a fabric in which forms a well-developed girdle while still retaining the strong maximum. Kink bands in the olivines are common, particularly in the larger grains (Fig. 16). The grain boundaries tend to be straight •and to have triple grain boundary-angles of 120°. The olivine from the Nicola Lake nodule has a very weak prefered orientation in which the three principal optic directions are mutually perpendicular. Each maximum is ill-defined and the fabric is essentially random (Fig. 17). Kink bands in these olivines are also fairly common but polygonisation is not a textural feature of the olivines of this nodule. The texture of this nodule, and also others of this suite, is typical of cumulate rocks in which the grain boundaries are mutually interfering (Fig. 18). Figs. 19, 20 and 21 are the fabric diagrams for the olivines of the Castle Rock nodules. The main feature of these is that b' forms a strong maximum perpendicular to a girdle. °< and (3 form less prominent maxima within the girdle. Specimen Explanation of Figs. 13, 14, 15 and 1? Olivine fabric diagram of specimen JL-A from Jacques Lake. 50 grains. Contours at 2, 4, 6, 8$ of 1$ area. . Maximum concentrations are 8,8,10$ for , /3, 0 respectively. Mode: ol. 77. opx. 11, cpx. 11, spin. 1. Fig. 14 Olivine fabric diagram of specimen JL-50 from Jacques Lake. 50 grains. Contours at 2, 4, 6, 8$ of 1$ area. Maximum' concentrations are 6, 10, 10$ for ©<; [3 / 5 respectively. Mode* ol. 63, opx. 19. cpx. 17, spin. 1. Fig* "1? Olivine fabric diagram of specimen JL-24 from Jacques Lake. 50 grains. Contours at 2V 4,':6,^8,1,10$' of 1$ area. Maximum concentrations are 14, 8, 10$ for <* } /3 , X respectively. Modes ol. 90, opx. 2, cpx. 7» spin. 1. Pig* 17 Olivine fabric diagram of specimen NL-8 from Nicola Lake. 50 grains. Contours at 2, 4, 6,,,8$ of 1$ area. Maximum concentrations are 8, 8, 10$ for ot f3} ^ respectively. Modei ol. 83, opx. 7» cpx. 9» spin. 1. ' ' Fig. 13 Fig. 14 Fig. 16 Strained olivine with abundant kink bands; specimen JL-50 from Jacques Lake. X-30j crossed nlcols. Fig. 18 Mutually interfering grain boundaries between olivine and orthopyroxene (grey and yellow); specimen NL-8 from Nicola Lake. X-30; crossed nicols. ,26 S-4 is layered and has been described previously. The orientation of the layering is shown in Pig. 20. The plane of the layering and the ^3 - <* girdle coincide and X is perpendicular to the layering. Many of the olivines of the Castle Rock nodules have kink bands. These are found only in the larger grains. The smaller ones are strain-free and form a mosaic with straight grain boundaries meeting at 120°, (Fig. 22). In addition to the olivine, an enstatite fabric diagram for sample CR-8 was constructed (Fig. 23). The enstatite in this nodule has a poor prefered orientation. There is a suggestion that y is perpendicular to an ill-defined ^" °< girdle, but more data are required to confirm this. Comparison with the olivine fabric from the same rock (Fig. 21) shows that there is apparently no correspondence between the olivine and the enstatite fabrics. This in in contrast to the findings of Collee (1963) and Rodgers and Brothers (1969). (c) Discussion. According to Rodgers and Brothers (1969) there are five orientation rules for olivine in ultramaflc nodules. These are t (1) o< maximum perpendicular to a f3 ~Y girdle. (2) X maximum perpendicular to a ~°< girdle. (3) V and f3 maxima perpendicular to a cx girdle. (4) ft , ) °* mutually perpendicular. (5) No obvious orientation. Transitional fabrics also occur. The fabrics which are found in lherzolite nodules are (1) and (4). This study has shown that .27 Explaination of Figs. 19. 20, 21 and 23. Fig* 19 Olivine fabric diagram of specimen S-2 from Castle Rock. 100 grains. Contours at 1, 2, 4, 6, 8, 10$ of 1$ area. Maximum concentrations are 6, 7, 12$ for <=< , /3} a respectively. Modej ol. 77. opx. 16, cpx. 6, spin. 1. Fig. 20 Olivine fabric diagram for specimen S-4 from Castle Rock. 50 grains. Contours at 2, 4, 6, 8, 12$ of 1$ area. Maximum concentrations are 6, 12, 18$ for^/^,0 respectively. Mode: ol. 72, opx. 15. cpx. 12, spin. 1. The great circle is the plane of the layering. Fig. 21 Olivine fabric diagram for specimen CR-8 from Castle Rock. 50 grains. Contours at 2, 4, 6, 8$ of 1$ area. Maximum concentration are 8,8,10$ for °<, fi\ 0 respectively. Mode: ol. 63, opx. 23, cpx. 12, spin. 2. Fig- 23 Enstatite fabric diagram for specimen CR-8 from Castle Rock. 50 grains. Contours at 2, 4, 6, 8$ of 1$ area. Maximum concentrations are 6, 6, 8$ for <=><} ft > 0 respectively. Mode: ol. 63, opx. 23, cpx. 12, spin. 2. Fig. 19 Fig. 20 ..: o29 Fig. 21 Fig. 23 .30 .31 olivine in lherzolite nodules may also follow rule (2). This type of fabric has only been found previously (apart from peridotites) in dunite and harzburgite nodules (Rodgers and Brothers 1969) and in an isolated "olivine nodule" (Talbot et. al. 1963). The mode of this nodule was not reported. Raleigh (1968) and Carter and Ave»Lallement (1970) have found that deformation of olivine at high temperature results in gliding on-^Okl^ . The axis of external rotation is [lOO^ . The orientation of the glide plane is dependant on temperature and strain rate. At high temperature and low strain rate the glide plane is (010). This is the plane on which slip occurs in most naturally deformed olivines, resulting in the formation of kink bands. Polygonlsation and recrystall!satlon result in the disappearance of the kink bands. Kink bands in olivine are not in themselves evidence of deformation in the solid state. Brothers (1962) has found that these bands are present in olivines in some basalts and gabbros which are undeformed. They may result from flow in the magma from which the olivine crystallised. Both kink banding and polygonlsation are common in the olivines of the Jacques Lake and Castle Rock nodules (Figs. 16 and 22). This suggests that these nodules have been deformed and have recrystalllsed (Raleigh 1968; Ragan 1969; Carter and Ave'Lallement 1970). The Nicola Lake nodules , on the other hand, are apparently undeformed since evidence of recrystallisa-tion is absent. Evidence from the olivine fabrics supports the above suggestions. The typical fabrics developed by olivine formed from a magma are types (4) and (5) and less commonly type (1) of the above .32 classification (Brothers 1959» 1962, 1964). The Nicola Lake olivine fabric is transitional between types (4) and (5) and could therefore have resulted from the olivine having crystallised from a magma. There is a range of fabrics shown by the Jacques Lake nodules. The fabric for JL-50 (Fig. 15) corresponds to type (4). JL-A (Fig. 13) is a modified type (4) fabric in which {3 forms a weak girdle. Slip on the system _|"okl|, [looj may explain the variations in the fabrics of this suite. If the initial fabric of the source rock was a simple pattern (possibly produced by crystal settling or perhaps by deformation) represented by JL-50 (Fig. 14), then deformation would result In the maxima for « Co/oJand =• being rotated about L'0olwhich is the zone axis for glide. This would result in the formation of girdles in <=< and (3 ; X would remain as a maximum. Specimen JL-24 (Fig. 15) shows this to forms a weaker girdle. Perhaps translation as well as rotation same plane as ex . An exact geometrical analysis of these fabrics is not possible since a sample of 50 grains is insufficient to determine precise angular relationships. Further complications are Introduced by the fact that the orientation of recrystalllsed olivine grains is,,in part, controlled by the Initial fabric (Ave»Lallement and Carter 1970). Despite the speculative nature of the above discussion, the range of fabrics (which are unlike those produced solely by crystal accumulation), coupled with evidence of recrystallisation given above suggests that the Jacques Lake nodules have been some extent where forms a girdle perpendicular to was involved is not in the .33 deformed. Recrystallisation is not complete as some of the larger olivines are still strained (Pig. 16). This could explain why there is a range of fabrics of the olivines of this suite. A single fabric type has been found in the Castle Rock nodules. This is type (2) of the above classification ( girdle). Fabrics developed by olivine formed from a magma have been reported above. As far as could be determined from the literature the fabrics of these olivines have no analogues among olivines from igneous environments. This type of fabric has been found in deformed Alpine peridotites and other types of ultramafic nodules (Ave*Lallement and Carter 1970; Talbot et, al, 1963; Rodgers and Brothers 1969). More commonly the typical olivine fabric of a strongly deformed peridotlte is one where oCX girdle (type (1) of the above classification). The nodules from Castle Rock are layered. This has been described in Chapter 3* Mineralogical layering in a rock can arise in at least two ways. These arei (a) by accumulation of minerals in a magma, and (b) by metamorphism and deformation or metasomatism. Both these processes may result in a prefered orientation of some mineral (in this case„;. olivine). To produce a fabric by crystal accumulation requires a dimensional orientation of olivine. No such orientation was found in the Castle Rock olivines. This can be explained by : suggesting that intercumulus growth took place which masked the original shape of the grains (Jackson 1961). However this would still produce a fabric which is typical of cumulate rocks. This is not the case for the Castle Rock nodules. .34 Metamorphic banding in peridotites is a common feature (Thayer 1969? Loney 1970). The banding takes the form of monomineralic veins and dykes and discontinuous stringers of minerals. Both types are found in the Castle Rock nodules (Figs. 8 and 9). The form of the spinel in specimen S-2 in which it appears to be flattened and to cut through the silicates (Fig. 9) suggests that there may have been an element of shear in the formation of the layering. Loney et. al. (1970) have found that mineralogical layering is sub-parallel or parallel to an axial plane foliation in a deformed peridotite. The axial plane is perpendicular to an maximum and parallel to a (Z-X girdle. It is possible that the fabric of these nodules is related to 1 a similar deformational feature although in this case the layering contains a (I girdle and X forms a maximum at right angles to it. This type of fabric Is characteristic of some deformed peridotites (Ave»Lallement and Carter 1970). It Is thus suggested that the Castle Rock and Jacques Lake nodules have been deformed. The basis of this is a comparison with the fabrics of deformed peridotites and magmatlc olivines and by the textures of the olivines in these nodules. Deformation must have occured prior to the nodules having been captured by the magmas which brought them to the surface. The different fabric types of the two suites suggests that the source rocks of the suites underwent different deformational histories. Factors which influence the type of fabric which is developed by the olivine of a peridotite are the orientation of the principal stresses with respect to the original fabric, .35 temperature, strain rate and the presence or absence of water (Ave'Lallement and Carter 1970). Pressure has apparently little influence on the fabric type although Ave»Lallement and Carter (1970) imply (p.2214) that very high pressure is required to -produce a fabric similar to that found in the Castle Rock nodules. This fabric type has.not been reproduced experimentally so that the conditions under which it formed are unknown. Experimental recrystallisation of olivine takes place at temperatures above 900°C (Ave»Lallement and Carter 1970) so that the deformation of these nodules probably occured at high temperature. .36 CHAPTER 5 Mineral Compositions. (a) Introduction. Partial mineral analyses presented in this study were carried out by means of atomic absorption spectrophotometry, electron microprobe and by wet chemical means. Details of the analytical methods and the precision of the results are given in Appendix 1. Only the four primary minerals (olivine, orthopyroxene, clinopyroxene, spinel) were analysed. The results of this study reaffirm the findings of many workers that the minerals of ultramafic nodules (particularly lherzolite) have a restricted compositional range. The compositional range of these minerals is similar in both major and minor elements to the ranges of other published analyses. This study also shows that the compositional range of the minerals in any particular suite of nodules is characteristic of that suite. The analysed minerals were chosen from nodules which had as wide a range of modal compositions as possible so as to include all possible variations. TABLE 1 j Modes of nodules with analysed minerals. Sample Locality 01. Cpx. Opx. Spin. JL-A Jacques Lake 77 11 11 1 JL-39 it 71 16 12 1 JL-10 ii 62 12 24 2 JL-B •i 68 22 8 2 JL-55 ii 80 8 10 2 95 Castle Rock 74 6 19 1 s-3 ii 83 5 11 1 CR-8- •i 63 12 23 2 ERC-11 H 66 17 15 2 NL-8 Nicola Lake 83 9 7 1 .37 (b) Olivine. Partial chemical analyses of olivines are given in Table 2. Of the four primary minerals, olivine has the most restricted range. It ranges from Fogo^ to FOCJI.Q *n tne Castle Rock and Jacques Lake nodules. The olivine of the Nicola Lake nodule has a composition of Fo^1#g. The range is similar to that found elsewhere (e.g. White 1966). There is no significant difference in the range of major elements in olivine among the suites. The range of values for Ni, Mn, Co, Cu and Zn and also CaO and Na£0 is similar to the range found elsewhere (Ross et. al. 1954; Forbes and Banno 1966; Simkin and Smith 1970; Carter 1970). The Pb content of the olivines is variable and appears to be high for ultramaflc rocks. Goles (1967) suggests 0.5ppm. as an average for ultramaflc rocks as a whole. While no data are available on the Pb content of olivines from ultramaflc rocks, this figure suggests that the olivines of these nodules are enriched in Pb (Table 2) relative to the olivines of other ultramaflc rocks. (b)Ort'hopyroxenes. Partial analyses of orthopyroxenes are given in Table 3. The Jacques Lake orthopyroxenes range in composition from Engo^zj, to En90.1 an<* tn°se from Castle Rock range from En^o.5 to En9i.9« The Nicola Lake orthopyroxene'has a composition of Eng^^. The Castle Rock samples are slightly more magnesian than those of Jacques Lake or Nicola Lake. The AlgO^ content ranges from 3«50 to 5.90$. With one exception (ERC-11), the Castle Rock ortho-pyroxenes contain less AI2O3 than those of Jacques Lake (Table 3). TABLE 2 Partial chemical analyses of olivines. Sample 95 S-3 CR-8 ERC-11 Oxide(wt.$) FeO* 9.25 8.45 9.56 9.60 MgO 50.0 50.0 46.9 48.1 CaO 0.05 0.05 0.03 0.04 MnO 0.11 0.10 0.10 0.12 NiO 0.32 0.31 0.32 0.34 Na20 0.21 0.19 0.22 0.08 Fe/Fe+Mg 10.13 9.20 10.29 10.08 Minor elements in ppm. Nl 2472 2845 2511 2640 Mn 897 751 763 911 Co 119 133 122 140 Zn 53 52 49 62 Pb 80 64 28 64 Cu 1.2 0.8 0.6 0.7 *Total iron as FeO NL-8 JL-A JL-39 JL-10 JL-B JL-55 8.38 9.66 9.91 9.95 9.70 10.42 >2.8 48.4 47.9 47.1 50.8 48.3 0.04 0.05 0.03 0.04 0.05 0.04 0.09 0.12 0.12 0.11 0.12 0.16 0.24 0.33 0.35 0.33 0.31 0.36 0.19 0.33 0.6l 0.32 0.59 0.27 8.18 10.08 10.39 10.22 9.71 10.52 1895 2619 2737 2604 2465 2861 702 929 956 889 913 1027 132 140 129 125 130 137 46 48 62 55 57 48 22 35 78 80 77 72 1.1 1.2 1.2 1.2 1.0 0.7 00 TABLE 3 Partial chemical analyses of orthopyroxenes. Sample 95 S-3 CR-8 ERC-11 NL-8 JL-A JL-39 JL-10 JL-B JL-55 Oxide(wt.$) • TiO? 0.06 0.06 0.08 0.14 0.10 0.22 0.26 0.14 0.16 0.12 Al?0o 3.68 3.54 3.50 5.90 4.04 4.46 4.68 4.40 4.60 4.66 CroOi O.37 0.35 0.31 0.26 0.10 0.46 0.48 0.48 0.33 0.33 Fe0*J 5.80 5.27 5.74 5.91 6.17 6.43 6.42 6.17 6.20 6.56 MgO 32.2 33.6 32.5 31.6 29.5 31.9 31.3 30.7 31.6 31.1. CaO 0.76 0.73 0.67 1.33 1.03 0.94 0.82 0.85 0.79 1.3& MnO 0.11 0.11 0.11 0.11 0.12 0.11 0.10 0.11 0.11 0.11 N10 0.11 0.11 0.12 0.12 0.12 0.11 0.10 0.11 0.10 0.12 Na20 0.21 0.13 0.27 0.45 0.24 1.00 0.32 0.31 0.24 0.50 K20 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 Fe/Fe+Mg 9.20 8.08 9.02 Fe/Fe+Mg+Ca 9.06 7.99 8.90 Mg/Fe+Mg+Ca 89.42 90.60 89.76 Ca/Fe+Mg+Ca 1.52 1.4l 1.34 CaTs(mol.$) 3 3 2 MgTs(mol.#) 2 4 3 Minor elements in ppm. Tl 330 36O 480 Cr 2500 2380 2130 Ni 838 856 889 Mn 839 743 812 Co 63' 60 63 Zn 40 33 3^ Pb 22 27 23 Cu 1.2 1.1 1.1 9.51 10.51 10.16 10.31 10.15 9.91 10.59 9.25 10.28 9.97 10.14 9.94 9.75 10.18 88.09 87.51 88.16 88.19 88.26 88.66 87.06 2.66 2.21 1.87 I.67 1.80 1.59 2.76 5 0 2 2 3 3 5 7 0 5 4 2 4 4 840 600 1320 1560 840 960 720 1750 675 3130 3250 3250 2250 2250 906 934 851 820 815 761 931 823 849 854 882 - 862 866 903 70 64 66 68 65 59 64 40 38 38 36 38 46 33 42 33 29 39 33 49 27 1.1 0.6 1.1 0.5 1.0 1.9 2.0 *Total iron as FeO .40 The above compositions are similar to the compositions of orthopyroxenes from other suites (e.g. White 1966). The range of values for Ti, Cr, Ni, Mn, Co, Cu, Zn and also Na20, K20 and Cao is similar to that found elsewhere (Ross et. al. 1954; White 1966; Carter 1970). There is a difference between suites in some of the trace element contents, independent of variations in major element concentrations. This is discussed in Chapter 7. As with the olivines, the orthopyroxenes have a high Pb content relative to the average for ultramaflc rocks (Table 3). (d) Clinopyroxenes. Partial chemical analyses of clinopyroxenes are given in Table 4. In terms of three end-members the Jacques Lake clino pyroxenes range in composition from Di^^^En^^^^Fs^^^ to Di48.4En46.6Fs5.0' those from Castle Rock range from Di#0^58.6Fs5.i1. to Di4-p>'pEn48,8Fsi4.,3. The Castle Rock clino pyroxenes are generally more magneslan and less calcic than those from Jacques Lake (Table 4). The Nicola Lake specimen is the least calcic with a composition of Dlii,^tjEn^itQFs^ti^, The clinopyroxenes contain from 2.78 to 7.12$ AI2O3. With one exception (ERC-11), the Castle Rock clinopyroxenes contain less AI2O3 than those of Jacques Lake. The Nicola Lake specimen has an intermediate concentration of AI2O3. The AI2O3 content . increases with Na20 and TiC>2 (Fig. 24) which suggests that these elements are substituting for Al in the pyroxene structure. The above compositions are similar to the compositions of clino pyroxenes of other lherzolite suites (e.g. White 1966). TABLE 4 Partial chemical analyses of clinopyroxenes. Sample 95 S-3 CR-8 ERC-11 NL-8 JL-A JL-39 JL-10 JL-B JL-5: Oxlde(wt.#) TiC-2 0.20 0.05 0.22 0.40 0.40 0.45 0.64 0.48 0.60 0.52 A1203 4.92 2.78 5.00 6.70 5.90 5.60 6.84 5.60 7.12 7.10 Cr20o 0.88 0.93 0.85 O.83 0.85 0.90 0.85 0.83 0.80 0.90 FeO*^ 2.44 2.27 2.26 3.40 2.99 2.75 2.69 2.74 2.61 2.74 MgO 15.^ 18.2 16.4 20.8 15.8 15.6 14.6 15.5 14.3 14.4 CaO 20.0 19.6 21.8 19.6 20.8 20.9 21.3 20.0 17.8 18.8 MnO 0.06 0.06 0.05 0.08 0.07 0.07 0.07 0.07 0.07 0.07 N10 0.06 0.07 0.06 0.08 0.08 0.06 0.05 0.05 0.05 0.05 Na20 1.00 0.63 1.26 1.09 1.39 1.00 1.54 0.90 2.11 1.34 K20 0.01 0.01 0.01 0.01 0.01 0,01 0.01 0.01 0.01 0.01 Fe/Fe+Mg 8.19 6.54 7.18 8.55 9.60 9.02 9.37 9.02 9.31 9.66 Fe/Fe+Mg+Ca 4.32 3.66 3.96 5.37 5.41 4.81 5.00 4.69 4.91 4.97 Mg/Fe+Mg+Ca 48.40 52.26 51.20 58.55 50.97 48.16 48.19 48.26 48.66 47.06 Ca/Fe+Mg+Ca 47.28 44.08 45.12 35.26 43.26 46.91 46.63 47.93 47.24 48.52 CaTs(mol.Jg) 3 6 4 6 5 6 4 9 6 10 Minor elements in ppm. Ti 1200 300 1320 2400 2400 2700 3840 2880 36OO 3120 Cr 5990 633P 5830 5650 5820 6160 5820 5650 5^70 6160 Ni 482 5^7 497 615 608 440 398 413 361 357 Mn 488 475 413 638 529 540 563 558 542 562 Co 50 50 46 81 63 49 48 50 53 42 Zn 21 20 18 26 22 19 19 20 15 16 Pb 89 83 69 60 104 84 77 73 10b 77 •Total iron as FeO 7 O AA O O 4' 3' 2-5 O O % Na20 AZ Fig. 2 4 : Variation of Na20 and Ti02 with Al203 in clinopyroxenes. A Jacques Lake, o Castle Rock. • Nicola Lake. 7 -O A A 55 O O A A 3 2-5 O O •3 % Ti©2 .^3 The composition of the clinopyroxene is one means of distinguishing nodules of the lherzolite series from those of the dunite-wehrlite-gabbro series and the eclogite series (White 1966; Kuno 1969). Consequently the variation of AI2O3, CaO and Na20wlth the ratio Mg0»PeO has been plotted in Pig. 25. As can be seen all the cllnopyroxenes from this study fall in the field of the lherzolite series. The range of values for Ti, Cr, Ni, Mn, Co, Cu, Zn and also Na20 and K2O are similar to the values reported elsewhere (Ross et. al. 1954; White 1966; Carter 1970). There are differences between suites in some of the trace element contents, independant of variations in major element concentration. This is discussed in Chapter 7. As with the olivines and ortho-pyroxenes, the clinopyroxenes have a high Pb content (Table 4) compared to the average given by Goles (1967) for ultramafic rocks, (e) Spinel. Partial chemical analyses of spinels are given in Table 5« The composition of natural chrome-bearing spinels adheres closely to the model formula (Mg,Fe2+) (Cr,Al,Fe:3+)20^; the sum of the oxides is generally more than 9&% (Irvine 1965). Consequently only these constituents were analysed to determine the variations in composition. The composition of chromian spinels may be graphically represented by means of a triangular prism, each of the six corners corresponds to one of th;e end-members (MgCr20^ ; FeCr20^s MgFe204; Fe203? MgA^O^jFeA^Oj^). Plotting is done by 8 7 6 i 3 5 CM 3 2 I 22 2! ^20 o . 18 17 Lherzolite series O O A ."44 O O o Wehrlite series O Lherzolite series O Wehrlite series o O 9 Fig. 25 Lherzolite series O O A Wehrjite series 8 7 6 MgO •. FeO Variation of ALp3, CaO and ISlOgO with MgO: FeO in clinopyroxenes. The lines are the boundary lines between the compositional fields of lherzolite and wehrlite series clinopyroxenes;. taken from Kuno (1969). A Jacques Lake: o Castle Rock: Nicola Lake. TABLE 5 Partial chemical analyses of spinels. Sample 95 Oxide(wt.%) MgO 20.27 FeO 8.91 Fe203* 2.98 AloOo 50.49 Cr203 15.38 Total .98.03 , Fe2*/Fe2++Mg 19.8 F e 3+/F e^+Al +C r 3 • 0 Al/Fe<>Al+Cr 79.0 Cr/Fe-' +Al+Cr 18.0 •Calculated by assuming model formula RR20i}> Ionic formula based on 32 (0) MgOJ. 6.507 6.55^ 6.671 6.760 5.909 6.052 5.900 6.275 6.391 5.668 Fe?t 1.605 1.694 1.476 0.789 2.007 1.886 1.885 2.002 I.563 1.815 Fe-* 0.729 0.852 0.498 1.335 - 0.068 0.144 Al 12.822 9.328 13.167 13.353 14.401 13.940 14.728 13.777 14.736 15.194 Cr 2.620 5.656 2.237 1.415 1.655 2.102 1.463 1.965 1.151 1.151 Rot 8.112 8.248 8.14? 7.5^6 7.915 7.940 7.755 8.225 7.95^ 7.483 RJ 15.926 15.836 15.902 16.103 16.056 16.042 16.164 15.801 16.031 16.345 s-3 CR-8 ERC-11 NL-8 JL-A JL-39 JL-10 JL-B JL-55 19.56 9.01 5*03 35.19 31.82 21.38 8.43 3.I7 53.35 14.19 23.58 4.43 8.42 54.19 8.43 19.0.5 11.53 58.71 10.06 19.73 10.95 57.39 12.93 20.15 11.29 63.57 9.24 19.93 11.33 0.46 57.35 11.76 21.86 9.53 0.97 63.72 7.42 19.06 10.88 64.59 7.30 100.61 100.52 99.05 99.3^ 101.00 104.25 100.93 103.30 IOI.83 20.5 5.2 56.5 38.3 18.1 3.1 80.8 16.1 9.5 8.2 82.2 9.6 25.3 88.6 11.4 23.7 85.5 14.8 24.2 90.2 9.8 24.2 0.4 86.3 13.4 20.3 0.9 91.9 7.2 24.3 92.2 7.8 .46 projection as shown in Fig. 26. The compositions of the analysed spinels are shown on the two projections. As can be seen, the spinels fall close to the MgAl20^ apex of the prism. This is similar to other analysed spinels from lherzolite nodules (Ross et. al. 1954; El Hamad 1963; Ishibashi 1969; Kutolin and Frolova 1970; Carter 1970). Spinels from other types of nodules do not necessarily plot in this part of the prism (Kutolin and Frolova 1970). Each nodule suite is characterised by having the composition of the spinel phase lying in a particular volume of the spinel prism (Fig. 26). The Castle Rock spinels are richer in Fe20^ and Cr203 than those from Jacques Lake and Nicola Lake; also the Cr20^ content of the Castle Rock spinels is more variable. The range of MgO variation Is small. The Nicola Lake nodule has the least magnesian spinel. There is no difference in the range of MgO content between the Jacques Lake and Castle Rock spinels. The relatively high Fe203 content of the Castle Rock spinels (Table 5) suggests that these crystallised under higher f02 than the Nicola Lake and Jacques Lake spinels (Irvine 1965. 1967). .47 Fig. 26 : Composition of analysed spinels. Jacques Lake -.o Castle Rock: •Nicola Lake, .48 CHAPTER 6 The Distribution of Iron and Magnesium  Between Coexisting Minerals. (a) Theory. The distribution of cations between coexisting phases has been discussed in detail by Ramberg and DeVore (195U. Mueller (1961), Bartholome (1962), Kretz (1961, 1963) and more recently by Grover and Orvllle (1969). The following summary is based on the work of Kretz (196l, 1963). An equilibrium exchange reaction for the distribution of A and B between coexisting phases (A,B)M and (A,B)N can be written: AM +BN = BM +..AN (1) The thermodynamic equilibrium constant for such a reaction is: aBM•aAN  KD = aAM * aBN where apQ is the activity of the appropriate end-member compound. Equation (2) can be written: YBM / BM YAN / AN B ,A B ,AA • A A KD ~ AM / AM' BN / BN ^> . AA . /\ A . . A B where XpQ is the mole fraction of P in PQ and A |Q is the , activity coefficient of P in PQ. If (A,B)M and (A,B)N behave as ideal solutions, then the activity coefficients are unity and equation (3) becomes: yBM YAN KD = AB A (l» YAM yBN A ,AB .49 If the Ideal solution model is correct then KD is a function of temperature and pressure only. The temperature dependence of Kp is given by: e) lnKD = £1H (5) ^ T / P RT2 where Z^H° is the change in enthalpy of reaction (1), R is the gas constant, and T the temperature in °K. The pressure dependence of is given by: / d lnKD^ _ _ £ vo (6) \ i p _/T RT where A V° is the molar change in volume of reaction (1) and P is the pressure in bars. As a first approximation, olivine, orthopyroxene and clino pyroxene can be treated as ideal solid solutions of the type (A,B)M and equation (4) can be used to determine the distribution coefficient with respect to the exchange of iron and magnesium between two coexisting phases. The distribution of iron and magnesium between spinel and any of the above minerals is considered later. The appropriate reactions 1S>J=H FeSiC-3 •+ MgSio.502 = MgSiC^ + FeSi0.502 (7) for olivine and orthopyroxene, and KD<1? = yopx yol (8) AFe ,AMg For coexisting pyroxenes the reaction is: .50 CaFeSl206 + MgS103 = CaMgSi206 + FeS103 (9) and KD(2> = X"g -XFe (10) opx cpx X .X Fe Mg For coexisting olivine and clinopyroxene the reaction is; CaFeSi206 + MgSig^Og = CaMgSi20£ + FeSl0>^02 and Ycpx vol KD<3) - „^x ' H (12) AFe ,AMg (b) Results for coexisting silicates. Values of KD(1), KD(2), and KD(3) are listed in Table 6. Also given is an analysis of the variance for the KJJ values between the Jacques Lake and Castle Rock minerals. The Nicola Lake minerals are considered separately. The Kp»s were calculated by assuming that all the iron in the minerals is in the ferrous state. While this is a good approximation for olivine and to some extent orthopyroxene, it is not so good for the clinopyroxenes since they may contain up to 2% ferric iron (Ross et. al. 995*0. Nevertheless it is still usefull to calculate the K^'s so obtained and compare the results for each suite. KQ(1) and KD(3) for the Jacques Lake nodules are less than the values for the Castle Rock nodules. This Is a significant difference as the F values exceed the 1% level of F (Snedecor and Cochran 19^7). Also, KD(1) and KD(3) for the Nicola Lake .51 TABLE 6 Values of Kp for coexisting olivine and pyroxenes, with analysis of variance. Sample Locality KD(D KD(2) KD(3) JL-A Jacques Lake 0.99 0.88 1.13 JL-39 •i 1.01 0.90 1.12 JL-10 •i 1.01 0.88 1.15 JL-B 0.97 0.93 1.04 JL-55 •I 0.98 0.91 1.11 95 Castle Rock 1.25 0.87 1.26 S-3 •i 1.16 0.79 1.45 CR-8 •i 1.17 0.78 1.48 ERC-11 ii 1/08 0.89 1.20 Variance Between suites .006 .008 .127 .016 .065 Within suites .002 Degrees of Between suites 1 1 1 freedom Within suifes 7 7 7 F 33.00 4.00 15.88 NL-8 Nicola Lake 0.89 0.90 0.84 sample are lower than any of the others. There is no difference in the values of KD(2) between any of the suites. The distribution coefficients for each mineral pair are similar to the coefficients determined from other lherzolite nodules (Kretz 1963s O'Hara 1963J White 1966). Direct comparison with other suites is not possible since total iron was expressed as FeO. In the ideal solution model, KD is a function of pressure and temperature only. If the minerals depart from ideality, then Kp will also be a function of composition, since the activity coefficients will not be unly. Nafzlger and Muan (1967) found that magnesian olivine is slightly non-ideal but ortho-.52 pyroxene is ideal. However, introducing activity coefficients ;V in equation (3) to evaluate will not change the relative values of KQ since there is a restricted range of compositions. It was found that each Kj) is independent of any other component in the minerals. Therefore the conclusions with respect to the differences in between the suites are still valid, (c) The effects of temperature and pressure. Medaris (1969) has determined experimentally the partition ing curve for iron and magnesium between olivine and ortho pyroxene at 900°C. The appropriate expression is: log [7^)01 = O.I630 + 1.1128log f^ll) opx (13) ^ XMg ' \ XMg I where Xpe and XMg are the mole fractions of iron and magnesium in the minerals. There is good agreement between this curve and the theoretical partitioning curve derived by Grover and Orville (1969)i who considered the exchange of iron and magnesium between olivine and orthopyroxene to take place between a single site in olivine (M^ and M2 are energetically equivalent) and a double site in orthopyroxene (M^ and M2 are energetically distinct). Equation (13) therefore appears to express satisfactorily the partitioning of iron and magnesium between these minerals at 900°C. Medaris also found that partitioning was not significantly temperature dependant between 900 and 1300°C. Despite the fact that temperatures of equilibration cannot be determined from the composition of coexisting olivine and orthopyroxene it is still useful/i to compare the distribution .53 of Iron and magnesium with the experimehtallydetermined curve in order to see if the minerals crystallised under equilibrium conditions. Pig. 27 is a plot of (XFe/XMg)ol versus (Xpe/XMg)opx on a logarithmic scale. The curved line is the partitioning curve determined by Medaris (1969). All pairs plot close to, but slightly below the curve. This is in agreement with other olivine-orthopyroxene pairs from lherzolite nodules (Medaris 1969). The points for both the Jacques Lake and Castle Rock pairs lie parallel to the curve which suggests that all the pairs crystallised under equilibrium conditions. The effects of pressure on the distribution coefficient has been given in equation (6). Using molar volume data on synthetic olivine and orthopyroxene Medaris (1969) found that Kj)(l) is not significantly dependant on pressure. However this conclusion is based on the assumption that A V° does not change with pressure or temperature. The change in volume for reaction (7) is given by: A V° " VMgSi03 + VFeSi0.502 ~ VFeSi03 " vMgSi0.502 (14) where V° is the molar volume of phase j. However the orthopyroxenes contain a considerable amount of Al (substituting for Si and (Mg.Fe)). Therefore the volumes which are used in equation (14) should be partial molar volumes and not the volumes of the pure end-member components. Since the Al content of the orthopyroxenes i:s different for each suite (Table 3) , it is to be expected that Av° will be different for each suite and hence the effect of pressure cannot be entirely disregarded. It is possible that the differ-'.'54 ~ 1 • • • • •• 11 " - i • • • • • — • ™ •• —»" ' •Ol o,i I.O Fe : Mg opx. 2+' Fig. 27 : Distribution of Fe and Mg between olivine and orthopyroxene. The curve is the equilibrium partitioning curve at 900°C determined by Medaris (1969). A Jacques Lake : o Castle Rock: • Nicola Lake. .55 ences in KD(1) between the suites may be due to differences in pressure. It is not possible to determine the direction of pressure change since the precision of the available data does not warrent the calculation. Kretz (I963) has shown that the distribution of iron and magnesium between coexisting pyroxenes is a function of temperature and is constant with respect to pressure. Kretz (1963) also showed that K^(2) increases with temperature for igneous and metamorphic pyroxenes and that the values for pyroxenes from ultramafic nodules tend to approach unity. The data from this study agree with the latter observation. The data suggest that all the pyroxene pairs crystallised within the same range of temperatures since there is no difference in the range of values of K^(2) between the suites. It will be shown in the next section that the Castle Rock nodules formed at a higher temperature than those of Jacques Lake, and that the Nicola Lake nodule formed at a lower temperature than the others. It might be expected therefore that the KD(2) values might show this. This is not the case. Therefore it is likely that some other factor besides temperature is affecting the values of Kp(2). This factor is possibly the AI2O3 content of the pyroxenes, although no direct relationship between the AI2O3 content of the pyroxenes and K^(2) could be found.' As with Kp(l), the pressure effect on Kj)(2) is dependant on the quantity AV°. The effects of AI2O3. on this are unknown, so that KD(2) could change significantly with pressure. Also,Z\H° may be affected by the Al20o, content of the pyroxenes. Therefore it is not possible to determine the conditions of formation of these minerals from the distribution of iron and magnesium between them. No experimental or empirical data on the effect of temperature and pressure on the distribution of iron and magnesium between coexisting olivine and clinopyroxene are available in the . literature. It is evident from Table 6 that iron and magnesium are distributed differently between these minerals in each nodule suite. It is assumed that this is a result of different conditions of formation. It Is not possible to evaluate variations in Kp(3) with respect to pressure and temperature using data for the iron-magnesium end-members since Al^Oo. in the clino-pyroxenes probably affects the appropriate thermodynamic functions, (d) The distribution of iron and magnesium between coexisting olivine and spinel. The distribution of iron and magnesium between coexisting olivine and spinel has been discussed in detail by Irvine (1965)* 2+ A reaction expressing the Mg-Pe exchange between olivine and spinel can be written; Fe2+Si0.502 •+ Mg(Crc<,Al^,Pe:^")20it = MgSi0#5O2 + Fe2+(Cr^,Al/9,Fe|+)20if (15) where are mole fractions of trivalent cations in the spinel and <x +/S^ = l. The thermodynamic equilibrium distribution coefficient for equation (15) lss a .SL* .a^ .a* KD(4) = MgSi0 502* FeCr20// FeAl^ FeFe20/j, ! 3 ~P * (l6) aFeSlQ ^ 502 • aMgCr20^'a MgAlgO^MgFe^ where is the activity of end-member j. This can be expressed in compositional terms by replacing a^ with (XJXJ) where A^ is the activity coefficient arid Xj is the mole fraction of the appropriate end-member. If ideal behavior is assumed, the activity coefficients are unity and equation (16) becomess oX ®P sp ^ i sp ^ xMg* ^xFeCr20i|,^ • ^FeAl20/4.) • (xFeFe204) KD(4) = oi sp sr-sp 77^ ? (17) AMg* VAMgCr20l+;' lAMgAl20^; * VAMgFe20^; Since x|P ~. = ( Fe ^ ^ / Cr \ =o(xsp-+ etc.; FECR2°4 iMg + Fe2*/ (cr + Al .+ Fe3+) APe2+ <K + {3 + X = 1, equation (16) reduces to K (4) = ^H2+ (18) x^.xJJP Fe Mg This is a result similar to that for coexisting silicates but is derived in this way to show the effects of the trivalent cations of the spinel on Kj-)(4). Table 7 gives the values of ^(4) calculated from equation (18). The method of calculating the uncertainties in KD(4) is given in Appendix 2. The Table also gives the analysis of variance in KD(4) between the Jacques Lake and Castle Rock suites, The Nicola La"ke pair is considered separately. As can be seen, the Nicola Lake sample has the highest KD(4). There is a TABLE 7 Values of Kp(4) for coexisting olivine and spinel, with analysis of variance. Sample Locality KD(4) lnKD(4) 95 Castle Rock 2.20 + .17 0.785 + .078 3-3 •I 2.55 .21 0.931 + .081 CR-8 ii 1.93 t" .16 0.651 .080 ERC-11 •• 0.93 + .10 -0.068 ± .037 JL-A Jacques Lake 2.77 ± .20 1.014 ± .071 JL-39 ii 2.78 ± .19 1.020 ± .069 JL-10 ii 2.81 ± .20 1.031 ± .071 JL-B •i 2.37 ± .19 0.860 ± .079 JL^555 •i 2.74 ± .19 1.004 ± .069 Between suites 1.39 Variance Within suites 0.20 Between suites 1 Variance Within suites 7 P 7.03 Nicola Lake 3.79 ± .30 1.330 ± .079 TABLE 8 Temperatures of formation for coexisting olivine and spinel. Sample Locality T C 95 Castle Rock 1632 ± 184 S-3 " 2214 ± 262 CR-8  1822 ± 231 ERC-11 " 10697 JL-A Jacques Lake 1133 ± 99 JL-39 " 1037 ± 90 JL-10  1113 ±98 JL-B " 1142 ± 122 JL-55  1002 ± 88 NL-8 Nicola Lake 838 ± 71 significant difference in the KD(4) values between the Jacques Lake and Castle Rock suites since the value of F exceeds the 5$ level of F (Snedecor and Cochran 19673. With one exception (S-3), all the Castle Rock KD(4)'s are greater than those of Jacques Lake. The distribution coefficient is related to the free energy of reaction (15) thus: lnKD(4) = and KD(4) = exp (zM^j (19) where A G = and Gj is the free energy of formation of the appropriate end-member in equation (15)» R is the gas constant and T the temperature of formation of the coexisting mineral pair. Substituting the free energy values given by Jackson (1969, p.64) (these are temperature dependant), taking R = 1.987. and solving for T, we have: 5580* + 10l8fi - 1720y + 2400 ' (20) T = 0.90* + 2.56(0 - 3.08 J - 1.47 + 1.9871nKD(4) Temperatures derived from equation (20) are given in Table 8. The method of calculating the uncertainty in each temperature is given In Appendix 2. The uncertainties listed in Table 8 represent the maximum possible errors in T which can be introduced as a result of analytical errors in the determination of Mg and Fe2+ in spinel and olivine and Cr, Al and Fe3+ in spinel. The error that may be introduced into the temperature values of Table 8 due to uncertainties in the free energy.values are much larger than the v6d; analytical errors. Possible maximum uncertainties in the free energies of the spinels alone could affect the temperature values by as much as ± 300°C but these uncertainties are not large enough to reverse the direction of reaction (15) and will have little effect on the relative temperatures (Jackson 1969). It is apparent from Table 8 that each nodule suite formed at a different temperature (within analytical error and assuming that all the minerals equilibrated at the same temperature). The nominal temperature of formation of the Nicola Lake suite is 838°C (assuming that sample NL-8 is representative of the suite). The nominal temperature of formation of the Jacques Lake suite is 1085°C (average of 5 temperatures). The nominal;'temperature.;; of formation of the Castle Rock nodules is more difficult to assess. There appears to be a range of temperatures but this may not be real as errors in K-Q(4) , and hence in lnKD(4), become increasingly important where KD(4) Is small since this term appears In the denominator of equation (20). As the derived temperatures seem unrealistically high the nominal temperature of formation of the Castle Rock nodules is taken to be l600°C. This is somewhat arbitrary but there can be little doubt that these nodules formed at temperatures much greater than those of Jacques Lake and Nicola Lake. The effect of pressure on KD(4) is that given in equation (6). Irvine (1965) has shown that the pressure effect is negligible at moderate pressues. Data to evaluate K^(4) at high pressure is lacking, but any variation in Kp(4) due to pressure is not likely to reverse the relative temperatures since any volume change will be small compared to the enthalpy change of equation (5). .61: (e) The distribution of iron and magnesium between coexisting spinel and pyroxenes. Reactions equivalent to (15) can be written for orthopyroxene and;;'.clinopyroxene: PeS103 + Mg(Cro(.,Al/9 ,Pe3+)20^ = MgSi03 + Fe( Cr^ ,A1^ .Fe^O^ (21) and CaFeSi206 + Mg(Crc< .Al^ tFe^)20^ = CaMgSigO^ + FetCr^.Al^.FeJ+JgO^ Analogous expressions for the distribution coefficients are: X§gX.Xp£2-f KD(5) = opx sp (23) xPe^.xMg and KD(6) = X^X.X|P2+ — (24) xcpx xsp ' Fe*' Mg Table 9 lists the values for Krj(5) and KD(6V for the analysed pairs and also the variation in these for the Jacques Lake and Castle Rock suites. There is no significant difference in these values between the suites. The values for ERC-11 are low in comparison to the others. This may indicate that the spinel in this nodule is not in equilibrium with the pyroxenes (and also with the olivine since KD(4) is low). Texturally, however, the spinel appears to be in equilibrium, so that the anomalous Kj)»s may be the result .62 TABLE 9 Values of KD(5) and KD(6), with analysis of variance, Sample Locality KD(5) KD(6) JL-A Jacques Lake 2.74 3.14 JL-39 " 2.78 3.08 JL-10  2.84 3.32 JL-B " 2.32 2.48 JL-55  2.71 2.99 95 Castle Rock 2.44 2.76 S-3 " 2.93 3.71 CR-8  2.24 2.85 ERC-11 " 1.00 1.13 Between suites .631 .307 Within suites .313 .547 Variance Degrees of Within suites 7 7 freedom Between suites 1 1 F 1.96 .1.78 NL-8 Nicola Lake 2.89 3.12 of an imperfect analysis. As can be seen from Table 4, ERC-11 is the least close to the model formula. Expressions analogous to (20) and (6) can be used to evaluate the temperature and pressure effects on KQ(5) and KD(6). There is a considerable amount of AI2O3 in the pyroxenes. Therefore there will be an exchange of Al between the pyroxenes and the spinel which will undoubtedly Influence the temperature and pressure effects. This effect is not known. .63 CHAPTER 7 The Distribution of Trace Elements  Between Coexisting Silicates. (a) Theory. The partitioning of trace elements between coexisting minerals has been dicussed in detail by Mclntyre (1963). An exchange reaction between two solid phases (Cr,Tr)A and (Cr,Tr)B can be written: CrA + TrB = TrA + CrB (25) where Tr and Cr represent the trace element and the element for which it substitutes (the carrier element) repectively. A and B refer to that portion of the mineral which does not take part in the reaction. Cr and Tr have the same valence. The equilibrium constant for reaction (25) is aTrA.aCrB KTr = (26) aCrA»aTrB where a-j is the activity of end-member J in equation (25). Equation' (26) can be rewritten: ,Tr i Cr xTrA•XCrB•A TrA *ACrB ^Tr ~ Ar>-rA,ATP-rR*'Vr.T'A,A' (27) CrA TrB CrA TrB where is the concentration of end-member j and / ^ is the activity coefficient of i in j.. If ideal behavior is assumed, then equation (27) becomes: K _ (XTr/xCr)A ""Tr (x /x ) -Tr Cr B •64 The distribution coefficient (K^r) is a function of temperature and pressure. The temperature dependency, given in equation (5) isJ I > lnKTr) _ ^o \ h T 'P RT2 where A H° is the enthalpy of reaction (25). R the gas constant and T is the temperature in °K. The pressure dependence, given in equation (6) is: / SlnK-r^ = _ Avo W P /T RT whereis the molar change in volume of reaction (25). P the pressure and the other symbols are the same as before. •Unfortunately An° and Av° for reactions involving trace elements are unknown so that trace element distributions cannot be used in geothermometry and geobarometry with any confidence. Nevertheless it is still useful to determine trace element distribution coefficients and anticipate that variations in these are due to vatiations:in the conditions of formation of the mineral pair in question, if it can be shown that the distribution coefficient is not dependant on the concentration of any other element. In olivine and pyroxenes the carrier element for divalent trace elements, may be either Mg or Fe . It is assumed that the M2 site in clinopyroxene is completely filled with calcium and that no exchange between a divalent ion and calcium takes place. The choice of Mg2+ of Fe2+ is not entirely arbitrary as stated by Matsui and Banno (1970). Burns (1970) has shown that .65 transition metal ions have a preference for certain sites in silicate structures. He has listed, among others, the following site preferences: olivine Ml 5 Ni2+,Co2+,Mg2+. olivine M2 ; Mn2+,Fe2+. orthopyroxene Ml ; Nl2+,Mg2+. orthopyroxene M2 ; Mn2+,Co2+,Fe2+. ] clinopyroxene Mi ; Ni2+,Co2+,Mn2+,Fe2+,Mg2+. Thus Mg2+ will be the carrier for Ni2+ in olivine and ortho pyroxene since both tend to occupy the same site in these minerals. Since both Mg2+ and Fe2+ occupy the Ml site in clinopyroxene the choice is not so clear. As there is a positive correlation between Ni,0 and MgO in the analysed is assumed that 2+ ? + Mg is the carrier for Ni in the clinopyroxemes also. Similarly 2+ ?.+ Fe is the carrier for Mn^and CoCT in all three silicates. It is assumed that all Co is in the divalent state in these minerals. Although "Zn is not a transition element it appears to behave as one in these minerals since it is distributed regularly among , the three silicates (Fig. 28). For this reason Zn is treated in the same way as Ni, Mn and Co ihfthis discussion. From size considerations it is likely that Zn2+ substitutes for Fe2+ in these minerals. (b) The distribution of Ni, Mn, Co and Zn. Since Ni, Mn, Co and Zn substitute for Mg or Fe in the olivine and pyroxene structure, the concentration of these elements is dependent on the concentration of the carrier element. In order to compare trace element concentrations, the ratio Tr/Cr .66 can be considered to be a measure of the trace element content of a mineral. (Ni/Mg)xl000, *(Mn/Fe)xlOO, (Co/Fe)xlOO and (Zn/Fe)xlOO for olivine, orthopyroxene and clinopyroxene are listed in Table 10. The variation in these ratios between the Jacques Lake and Castle R6ck-:"3uites Is examined by means of Snedecor's F test and the results given in Table 12. The Nicola Lake sample Is considered separately. The Nicola Lake olivine contains more Ni and Co than the other olivines and the pyroxenes contain less Mn and Co. The Jacques Lake and Castle Rock olivines contain similar amounts of these elements. The Jacques Lake diopsldes contain less Ni, Co and Zn than those of Castle Rock, and the enstatites less Mn and Co. These are all significant results since the value for F exceeds the 5% level of F in each case (Table 12). There is thus a fundamental difference in the concentration of some elements in the minerals of each suite of nodules. The minerals of each suite are characterised by a particular trace element "content". The distribution of Ni, Mn, Co and Zn between olivine and the two pyroxenes is regular. Fig. 28 shows the relative concent rations of these elements in the three silicates. The relative enrichmentfof these elements In coexisting olivine and pyroxene \3 1. Ratios involving iron were calculated with the assumption that all the iron is in the ferrous state. Generally the minerals (especially cllnopyroxenes) will contain some ferric iron. This will affect the magnitude of such ratios but the relative values will remain since there is a restricted compositional range which is assumed to apply to ferricciron alspa TABLE 10 (Tr/Cr) ratios of analysed minerals. .67 (Ni/Mg)xl000 ol. opx. cpx. Sample Locality JL-A Jacques Lake 8.98 4.42 4.69 JL-39 " 9.46 4.34 4.52 JL-10 " 8.81 4.41 4.42 .JL-B . • . 8.04 3.99 4.20 JL-55 " 9.91 4.97 4.12 95 Castle Rock 8.20 4.32 5.20 S-3 " 8.03 4.23 5.00 CR-8  8.89 4.54 5.02 ERC-11 " 9.09 4.75 4.91 NL-8 Nicola Lake 5.95 5.25 6.40 (Co/Fe)xl00 ol. opx. cpx. Sample Locality JL-A Jacques Lake .186 .132 .229 JL-39 " .168 .136 .230 JL-10 M .162 .136 .235 JL-B .172 .122 .261 JL-55 " .174 .125 .197 95 Castle Rock .161 .140 .263 S-3 " .202 .146 .284 CR-8  .164 .141 .261 ERC-11 " .187 .152 .307 NL-8 Nicola Lake .203 .133 .272 (Mn/Fe)xl00 ol. opx. cpx. 1.24 1.71 2.52 1.24 1.77 2.69 1.15 1.79 2.62 1.21 1.80 2.67 1.30 1.77 2.64 1.25 1.86 2.57 1.14 1.81 2.70 1.03 1.82 2.35 1.22 1.79 2.42 1.08 1.77 2.28 (Zn/Fe)xl00 ol. opx. cpx. .076 .076 .089 .071 .072 .091 .080 .079 .094 .O63 .095 .074 .061 .065 .075 .074 .089 .111 .079 .080 .114 .066 .076 .102 .083 .087 .098 .071 .079 .095 v6'8. cpx • Mn \ AZn \ . o Co / \ • Ni ° D n • i • ol opx Fig. 28 : Relative proportions of Mn, Zn, Co and Ni between coexisting olivine, orthopyroxene and clinopyroxene. a .69 is a function of size and site energy as predicted by Burns (1970). These enrichments are in agreement with the data from other ultramaflc rocks (White 1966; Mercy and O'Hara 1967; Carter 1970). The distribution of these elements between coexisting olivine, orthopyroxene and clinopyroxene is examined by means of the distribution function KTr (equation 28) where Tr = Ni, Mn, Co or En, Cr is the appropriate carrier element (Mg or Fe) and A and B are coexisting olivine, orthopyroxene or clinopyroxene. The results are given in Table 11. The variation in each KTr is examined by means of Snedecor's F test (Table 12). This applies to the Castle Rock and Jacques Lake suites onlyj the Nicola Lake results are considered separately. From Table 11 it can be seen that all the Nicola Lake KNi»s are less than the corresponding K^»s for the other two suites and that KN1(ol/opx) and KN1(ol/cpx) for the Castle Rock pairs are less than the corresponding ratios of the Jacques Lake suite. K^n(cpx/opx) for Nicola Lake is greater than the others;and for Jacques Lake /are- less than the Castle Rock coefficients. KCo (ol/opx) for Nicola Lake is greater than the Castle Rock and Jacques Lake coefficients. K^tol/cpx) for the Jacques Lake samples are less than those of the Castle Rock minerals. These comparisons are significant since the value of F exceeds the level of F in each case (Table 12). The conslstancy of the distribution coefficients within each suite suggests that the minerals are in equilibrium with respect to the elements which have been discussed. TABLE 11 Trace element dis KNi Sample Locality ol/opx ol/cpx JL-A Jacques Lake 2.03 1.19 JL-39 2.18 2.09 JL-10 II 2.00 1.99 JL-B II 2.02 1.91 JL-55 •1 1.97 2.38 95 Castle Rock 1.90 1.58 S-3 II 1.90 1.58 CR-8 II 1.96 1.71 ERC-11 II 1.91 1.85 NL-8 Nicola Lake 1.13 0.93 KCo Sample Locality ol/opx ol/cpx JL-A Jacques Lake 1.14 0.81 JL-39 II 1.24 0.73 JL-10 II 1.20 0.73 JL-B II 1.41 0.66 JL-55 1.39 0.88 95 Castle Rock 1.19 0.63 S-3 II 1.38 0.71 CR-8 •i 1.16 O.63 ERC-11 H 1.23 0.61 NL-8 Nicola Lake" 1.53 0.75 on coefficients. Mn opx/cpx ol/opx ol/cpx opx/cpx 0.94 0.73 0.49 0.68 0.96 0.70 0.46 0.66 1.00 0.64 0.44 0.68 0.95 O.67 0.45 0.67 1.12 0.73 0.49 0.67 0.83 O.67 0.45 0.67 0.83 0.67 0.49 0.67 0.91 0.57 0.44 0.77 0.94 0.68 0.50 0.74 0.82 0.61 0.47 0.78 KZn opx/cpx ol/opx ol/cpx opx/cpx 0.58 1.00 0.85 0.85 0.59 0.99 0.78 0.79 0.59 1.01 0.85 0.84 0.47 0.66 O.85 1.28 O.63 0.94 0.81 0.87 0.53 0.83 0.67 0.80 0.51 0.99 0.69 0.70 0.54 0.87 0.65 0.75 0.50 0.95 0.85 0.85 0.49 0.90 0.75 0.83 .71 TABLE 12 Analysis of trace element variance between Castle Rock and Jacques Lake suites. Ratio Variance Between Within suites suites Significance # (N.i/Mg)ol (Ni/Mg)opx (Ni/Mg)cpx (Mn/Pe)ol (Mn/Fe)opx (Mn/Fe)cpx (Co/Pe)ol (Co/Fe)opx (Co/Pe)cpx (Zn/Fe)ol (Zn/Pe)opx (Zn/Fe)cpx .031 .337 12.17 Not significant .002 .093 46.93 .918 .037 24.81 Significant at 0,5%' .010 .040 1.76 Not significant .010 .0014 7.14 Significant at 5% .030 .010 3.00 Not significant .0001 .0002 2.00 11 .0005 .00003 16.67 Significant at 1% .0052 .0005 10.61 Significant at 2,5% .00002 .00007 3.50 Not significant .00007 .00009 1.28 11 Significant at 0.5$ .0002 .00007 28.50 KTr Ni ol/opx Ni ol/cpx Ni opx/cpx Mn ol/opx Mn ol/cpx Mn opx/cpx Co ol/opx Co ol/cpx Co opx/cpx Zn ol/opx Zn ol/opx Zn opx/cpx •Degrees of freedom are 1 and 7 for between and within suites respectively #The significant levels of F are 16.24, 12.25, 8.07 and 5.59 at the 0.5, 1.0, 2.5 and 5.0$ levels respectively (Snedecor and Cochran 1967). .033 .004 8.25 Significant at 2.5$ .278 .029 9.59 Significant at 2,5% .033 .009 3.68 Not significant .007 .002 3.50 •1 .00005 .001 20.00 11 .006 .001 6.00 Significant at 5% .018 .010 1.80 Not significant .026 .006 4.80 11 .005 .002 2.32 ti .0002 .015 73.86 •1 .029 .004 6.86 Significant at 5$ .044 .097 2.20 Not significant .72 Although the distribution of trace elements among these minerals cannot be related quantitatively to the conditions of formation, it is clear that some elements are distributed differently among these minerals and that the observed distri bution cannot be related to different concentrations of trace or carrier element in the minerals. For example, it has been shown that Ni is distributed differently between olivine and enstatite in the Jacques Lake and Castle Rock suites, although the Ni content of these minerals in each suite is similar. Ni therefore appears to be sensitive to changes in environment. On the other hand, the Co content of the Jacques Lake pyroxenes is less than those of Castle Rock but the Co is distributed between the minerals in a similar way. Co is therefore not sensitive to changing conditions (at least in these minerals). The Nicola Lake pyroxenes contain less Mn than the other pyroxenes and is also distributed differently between these minerals. Each nodule suite is apparently characterised by different trace element behavior. The difference may be a difference in the content of some elements in the minerals or may be a differ ence in element distribution among the minerals which presumeably reflects different physical and chemical conditions at the source of the nodules. The above discussion Illustrates the importance of considering more than one element and pair of minerals when making inferences on the conditions of formation of. a series of rocks. While one set of distribution coefficients may not be significant with respect to changesof environment, the evidence of several sets .73 may reveal that suites of similar rocks have formed under different conditions, (c) Other elements. Other trace elements such as Cu, Pb, Cr and Ti are not amenable to a treatment such as given to Ni, Mn, Co and Zn as the position of these elements in the mineral structure is more uncertain and because incorporation of these elements in the structure involves a coupled substitution. Cu and Pb are distributed Irregularly among the three silicates (Pig. 29). Because of this, the distribution of these elements is considered no further. Tables 3 and 4 show that the concentration of Ti and Cr is greater in the clinopyroxenes than in the orthopyroxenes. This is in agreement with the data of White (1966). There appears to be no systematic difference between the suites in the Cr content of the pyroxenes. The exception to this is the Nicola Lake enstatite which is comparatively low in Cr. On the other hand the the pyroxenes of the Jacques Lake nodules have a higher Ti content than those of Castle Rock. The Nicola Lake pyroxenes have a similar Ti content to those of Jacques Lake. The distribution of Ti and Cr between the coexisting pyroxenes is examined by means of the Ne'rnst Distribution Law: TrA/Trfi = kTr (29) where Tr^ is the concentration of a trace element in phase j and krpr is a constant at any pressure and temperature. In this case A and B are orthopyroxene and clinopyroxene respectively. :?4 cpx / \ • Pb / \ # Cu / D \ /# \ / §/ • D • \ / D D» \ / ° • \ / t* \ ol opx Fig. 29 : Relative proportions of Pb and Cu between coexisting olivine, orthopyroxene and clinopyroxene. .75 TABLE 13 Values of kTi and kCr for coexisting pyroxenes with analysis of variance. Sample Locality kT1 kCr JL-A Jacques Lake 0.49 0.51 JL-39 0.41 0.56 JL-10 " 0.29 0.58 JL-B - 0.27 0.41 JL-55 " °»23 °«37 95 Castle Rock 0.28 0.42 y:? 1.20 0.38 CR-8 " 0.36 0.37 ERC-11 . " 0-35 0.30 NL-8 Nicola Lake 0.25 0.10 Between suites .097 «031 Variance n Within suites .080 .006 Within suites 7 7 Degrees of freedom Between suites 1 1 p , 1.21 5.17 The values for kTi and kCr are listed in Table 13. Also given is an analysis of the variance in kTr between the Jacques Lake and Castle Rock suites. The Nicola Lake pair are considered separately, As can be seen both kT1 and kCr are variable and there is no difference between the suites. The exception is the low kCr of vv the Nicola Lake pair. The reason for this is not known. The foregoing suggests that the distribution of Ti and Cr. between coexisting pyroxenes is not sensitive to different conditions of formation. The higher Ti content of the Jacques Lake pyroxenes is probably a result of a difference in the Ti content of the source rocks. .76 CHAPTER 8 The Origin of the Nodules. (a) Temperature and Pressure. On the basis of the distribution of Iron and magnesium between coexisting mineral pairs It has been shown that each suite of nodules formed under different P/T conditions. Variations in K-q reflect different temperatures of formation. Nominal temper atures of formation are 838°C for the Nicola Lake suite, 1085°C for the Jacques Lake suite and l600°C for the Castle Rock suite. While there is some doubt as to the absolute temperatures it is believed that each suite formed at different temperatures and that the relative temperatures are correct (Chapter 6). The effects of pressure on are unknown ,but:;ltils believed that different pressures are also responsible for variations in the distribution coefficients. This may be one reason for the comparatively high Castle Rock temperatures; the effects of pressure were not taken into account in the calculation of the temperatures. Al substitution in pyroxenes affects the values of KQ*S Involving pyroxenes (Chapter 6), but unfortunately quanti tative estimates of pressure cannot be made on the basis of variations in Kp. Al occurs in both sixfold and fourfold co-ordination in the pyroxene structure. This is a result of the requirements of charge balance. The appropriate subsltutions are Al*^ + A1V* for MgVI + SiIV and A1VI + NaVI11 for 2MgVI. If Al occurs in octahedral and tetrahedral co-ordination, then a Tschermak's component will appear, in an end-member calculation. Since only partial chemical .77 analyses were done, the amount of Tschermak's component can only be estimated. This was done by assuming that the weight percent of Si02 = 100 - j> Rx0y where Rx0y is the weight percent of any oxide. The calculation of the end-members was done by computer using the program PYREND (U.B.C. Dept. of Geology program; P.B. Read). The calculation of Si02 by difference and the fact that all iron was assumed to be in the ferrous state limits the accuracy .:: of the results. The error in estimating Si02 will be the sum of all the errors in the determined oxides. Because of this only the results of the calculation of the Tschermak's components are given (Tables 3 and 4), since these are the most significant with respect to pressureh(see below). CaTs is a component of all the clinopyroxenes. There appears to be no difference between the suites in the amount of CaTs in these pyroxenes. Both CaTs and MgTs are components of the ortho pyroxenes, except in the Nicola Lake ens-tafelt.e.,. Fig. 30 shows the proportions of CaTs, MgTs, diopside and enstati'tB In the pyroxenes. Since CaTs and MgTs are present in solution (except for Nicola Lake) in diopside and enstatite respectively it is reasonable to suggest that these minerals formed at high pressure(Boyd 1963; Kushiro and Yoder 1966; Kushiro 1969a). The stabilities of Al-diopside and Al-enstatite coexisting with spinel and olivine are shown In Fig. 32. Because the CaTs solubility in diopside Is complexly related to temperature, pressure, the amount of^Jadeite in the diopside and the nature and composition of the coexisting phases, it is not possible to make precise estimates of the pressures at which MgTs CaTs pyrope _^ grossular | Enstatite MgSi03 O Diopside CaSi03 Fig. 30 •. Composition of analysed, pyroxenes in terms of MgSi03- CaSiOj- AlgOj. opx. f + Jacques Lake, o Castle Rock. cpx. • Nicola Lake. A Jacques Lake, o Castle Rock. > Nicola Lake. .79 the diopsides formed,(Kushiro 1969a). Boyd and England (i960) and Boyd (1963) have shown that the AI2O3 content of enstatite coexisting with olivine and garnet increases with increasing temperature and pressure due to the coupled substitution 2A1 for (Mg + Si). Similar variations might be expected for enstatite coexisting with olivine and spinel. The A^O^ content of the analysed enstatites varies with that of the spinels and also between spinel and diopside (Fig. 3D* This suggests that the distribution of AI2O3 between spinel and pyroxene may be a function of pressure and temperature as is the case for co existing pyroxene and garnet. Unfortunately there are no experimental data to confirm this. The temperatures derived fr-omtthe distribution of iron and magnesium between coexisting olivine and spinel, and the high pressures inferred from the AI2O3 content of the pyroxenes are consistent with the experimentally determined field of spinel lherzolite. The stability fields of various ultramaflc mineral assemblages are shown in Fig. 321 taken from Green and Ringwood (1970). Spinel lherzolite has a fairly wide stability field which falls within upper mantle conditions. On the low pressure side of the lherzolite stability field plagloclase is stable and on the high pressure side garnet appears. The upper temperature limit of the stability of lherzolite is, of course, the lherzolite solidus. Many experimental studies have been carried out to determine the position of these boundary curves. Reactions between olivine and plagloclase to yield aluminous pyroxenes and spinel have been investigated by Kushiro and Yoder (1966). Such a reaction is: 70 60 H CL r so 40^ 30 70 60 Q. 50 40 30 O ,.80 O A A O O % Al203 cpx. AA O O O O 4 5 6 % Al2 03 opx. 7 7 Fig. 31 : Variation of Al203 between spinel and pyroxenes. A Jacques Lake : o Castle Rock •. • Nicola Lake. 8 a .81 50 Depth in Km. IOO 150 200 O lO 20 30 40 50 60 70 Pressure in Kb. Fig. 32 : Relative stabilities of various ultramafic mineral assemblages. (after Green and Ringwood 1970 ) .82 forsterlte + ahorthite = Al-diopside + Al-enstatite + spinel The exact position of the curve in natural systems is dependant on the compositions of the reacting phases which may vary. The relative stabilities of plagloclase and spinel bearing peridotites is,however,as shown in Fig. 32. Reactions defining the breakdown of spinel and the incoming of garnet are: enstatite -f spinel = forsterlte + pyrope and diopside + spinel = forsterlte + grossular (McGregor I967). There is no general agreement on the exact position of the, boundary curve (Green and Ringwood 1970). McGregor (1970) has shown that'the stabilities of spinel and garnet bearing peridotites is strongly dependant on the Cr20^yR2'C>3 ratio of the rock. >Spinel peridotites with a high Cr2C>3 J&203 rat^° are stable at higher pressures than those with a low 0X20^*112^3 ratio. It is to be expected that similar variations will also affect the lower limits of the stability of spinel lherzolite. At near solidus temperatures spinel lherzolites are stable to about 23Kb. depending on the ratio of trivalent oxides (McGregor 1968, 1970). Despite the uncertainties in the experimental data, the mineral assemblages and partial mineral analyses suggest that the nodules formed in the upper mantle where at near solidus temperatures, the pressure would be between 11 and 23 Kb. (equivalent to a depth between 35 and 75 Km.) (Fig. 32). Each nodule suite apparently formed under different conditions within the mantle. .83 Of the three suites, the Nicola Lake nodules formed at the lowest temperature and probably pressure. The enstatite of the noduleo from this suite contains no MgTs (Table 3) suggesting that most of the Al is in fourfold co-ordination and that this zxA? nodule formed at a lower pressure than those of the other suites* (Boyd 1963). The Jacques Lake nodules formed at temperatures well below those of Castle Rock and slightly above the Nicola Lake nodules (Table 8). The relative pressures are not so certain. The Jacques Lake pyroxenes contain more A^O-^ than those of Castle Rock (Fig. 3D but whether this is due to temperature, pressure or bulk composition is not known, A higher temperature does not necessar ily imply a higher pressure as the geothermal gradient in the Castle Rock region may be steeper than in the Jacques Lake region. However it is believed that the pressures at which the two suites of nodules formed were not the same. The basis of this inference is the different distributions of iron and magnesium between olivine and orthopyroxene for the two suites (Chapter 6). Supporting this is the different olivine fabrics in the suites (Chapter 4). The Castle Rock fabric may be due to relatively high pressure and temperature (Ave'Lallement and Carter 1970) but this requires experimental confirmation, (b) The nature of the source. The preceding discussion has.shown that the lherzolite nodules probably originated in the upper mantle. It is now necessary to decide which aspect of the mantle they represent. Two hypotheses .84 are considered. (a) They are crystal cumulates which precipitated from their present host rocks. (b) They are fragments of the mantle which may have been depleted by partial melting. If a cognate origin is proposed, it would be expected that there would be a wide range in the mineral proportions and compositions among a suite of lherzolite nodules. This is not the case for the Castle Rock or Jacques Lake suites. The range of compositions of the Nicola Lake suite is unknown. White (1966) and Kuno (1969) have shown that dunite, wehrlite and gabbro nodules have a wide compositional range and that the compositional trends of these nodules is distinct from those of lherzolite nodules. The variation in the wehrlite nodule series is thought to be due to crystal settling from a basaltic magma at depth. The narrow compositional range of lherzolite nodules is due to their being fragments of the mantle. Binns (1969) has found both lherzolite nodules and megacrysts of undoubted cognate origin in the same lava flow. These mega crysts which may occur as clusters-, include olivine, clinopyroxene, orthopyroxene and spinel and are quite distinct chemically (they are less magnesian) from the minerals of the lherzolites and «gf% appear to have originated at depth. It can be argued in this case that the lherzolites are not cognate but are residual mantle material and that the megacrysts represent the earliest crystal fraction of a basaltic magma produced in the upper mantle. Kutolin and Prolova (1970) and Aoki and Kushiro (1968) have examined .85 similar material and have come to the same conclusion. The petrofabric study has shown that the Castle Rock and Jacques Lake nodules have been deformed in the solid state. If these nodules are cumulates, then the following sequence of events might have taken place. Partial melting in the mantle occured. The liquid which was produced remained at depth while crystal fractionation took place. The cumulates which formed were then deformed before being brought to the surface. If this process did take place, it would require quiescent conditions in the upper mantle or lower crust to allow a crystal pile to accumulate. This is contrary to the conditions inferred from the olivine fabrics, and also contrary to conditions expected during partial fusion. On the other hand, a simple two-step process whereby fractional melting and ascent of the resultant liquid brought up fragments of the residual mantle rock would satisfy the requirement that the nodules have been deformed. The fabric Is then due to processes which operated prior to or perhaps during partial melting. It is possible that partial melting could result in the production of a basalt which remained at depth while precipitat ing crystals and forming a lherzolite. The magma could then have been removed and the lherzolite deformed. A second episode of melting could then have occured and fragments of the lherzolite caught up in the resultant basalt which brought them to the surface. Carter (1970) has examined lherzolites and other ultramaflc nodules from a single locality. He found that those lherzolites ("typical 4-phase nodules") which have a mode close to the olivine apex of the olivine-orthopyroxene-cllnopyroxene diagram (Fig. 3) generally have olivine with a composition more magneslan than Fo3g. Other lherzolites ("atypical 4-phase nodules") fall in the central part of the diagram and have olivines less magneslan than Fog^.These lherzolites appear to be undeformed and have cumulate textures, as opposed to the typical lherzolites which appear to r be deformed and recrysallised. Carter (1970) proposes a model whereby atypical lherzolites (and also wehrlites and pyroxenltes) are among the cumulates formed at depth from a basaltic magma. Typical lherzolites are probably residual products of partial melting, although in some cases an origin by accumulation cannot be discounted. This is similar to the model proposed by White (1966). Carter's (1970) modellisbas;€d on experimental work by Kushiro (1969b) and on an analysis of possible crystal-liquid paths during partial fusion according to the methods of Presnall (1969). White's (1966) model is based on petrography and mineral chemistry. Using the above model the' lherzolites of this study fall into the category of "typical 4-phase nodules" (Table 2, Fig. 3)• No wehrlites or pyroxenltes were found with the Jacques Lake lherzolites. A comparison with the above model and with work by Aoki and Kushiro (1968), Kuno (1969), Blnns (1969) and Kutolin and Frolova (1970) mentioned above suggests that the Jacques Lake nodules were not part of a cumulate series prior to their incorp oration into the Jacques Lake tuff. The full range of nodule types at Castle Rock is not known but since all the nodules which were studied fall into the category of "typical 4-phase nodule" and appear to have been deformed, antorigin by accumulation is discounted. .87 On the basis of the mode (Table 1) and the composition of the olivine (Fo^^>2) specimen NL-8 from Nicola Lake is a "typical" 4-phase nodule". The modes of the other nodules from this suite are variable (Fig. 3) although the full range is not known. The olivine fabric of this nodule and the textures of this and other nodules from this suite suggest that these nodules have formed by accumulation. The temperature at which the Nicola Lake nodules have inferred to have formed (nominal temperature is 838°C) is below any known lherzolite solidus, even under hydrous conditions (Fig. 33). The weight of the evidence therefore favours an origin by crystal settling and accumulation at depth for this suite. The nodules may be cognate with their enclosing basalt. One essential characteristic of a parental mantle rock is that it is capable of producing a basalt on partial melting at high pressure. There are two ways to consider this proposition. Firstly, direct experiments on the melting behavior of possible mantle rocks (natural or synthetic) can be made. Secondly, mixtures of basalt plus refractory residuum can be examined to determine .i;:-;^-;' whether these pairs result in a proposed mantle composition. Direct melting experiments on spinel lherzolite have been carried out by Kushiro (1968) and by Nishikawa (1970). The results of these studies and also of work by Kushiro (1969b) indicate that a silica undersaturated magma can be produced by partial melting of a spinel lherzolite. The type of magma which is produced depends on the degree of partial melting, on the presence or absence of water, and also the pressure at which melting took .88 place. The residue from such melting may be a more magnesian lherzolite, a harzburgite or a dunite. These may be produced under a wide pressure and temperature range and under both hydrous and anhydrous conditions. In additiom to the above experimental studies, work on the chemical relationships between basalt and ultramafic nodules has been done by Kuno (1969), Kuno and Aokl (1970) and Jackson and Wright (1970). They found that a basalt close to the composition of an olivine tholeiite could be produced from a pyroxene rich lherzolite, leaving a more magnesian lherzolite as residual material. Green and Rlngwood (1969) argue that alkali olivine basalt can be produced by a low degree {^20%) of partial melting of lherzolite at depths of 35-70Km. With an increasing degree of partial melting, olivine tholeiite can be produced. The above summary of recent work on basalt-lherzblite relation ships shows that lherzolite may be refractory residual material left after partial melting of primary mantle, which may itself be lherzolite (Kuno and Aoki 1970) or a mixture of lherzolite and garnet peridotite (Jackson and Wright 1970). Thus lherzolite nodules (including the nodules of this study) could be residual fragments of partially melted mantle. Lherzolite nodules are not generally considered to be parts of the primary mantle since they are too low in certain elements (K,Ti, P, Ba,Sr,Rb,Th,U and others) to produce basalt on partial melting (Harris et. al. 1967; Green and Ringwood 1969). The low K20 content of the minerals of these nodules supports this (Tables 3 and 4). .89 No chemical data on the enclosing rocks of the nodules In this study are avaiable. Nevertheless, in view of the close similarities in mineralogy, texture and mineral compositions of these nodules to nodules which have been studied in relation to their host rock chemistry, it is reasonable to suggest that these rocks are fragments of the mantle. A worldwide similarity of lherzolite nodules, regardless of the nature of the enclosing rocks, is an argument against lherzoliteimddules In general being cumulates. The temperatures at which theSCastle Rock and Jacques Lake nodules are; tnought^td have: formed are consistent with the hypothesis that the Castle Rock and Jacques Lake lherzolites are residual fragments of the upper mantle. The Jacques Lake nodules appear to have formed at 1085°C. This temperature is close to the solidus temperature of lherzolite at high pressure where PJJ2Q < ]?total and above the solidus temperature of lherzolite where PH2° = Ptotal at high Pressure (Kushiro et. al 1968) (Fig. 33). While the above temperature could be applied to the argument that the Jacques Lake nodules formed at a basalt llquidus temperature, the same cannot be said for the Castle Rock suite. The calculated temperatures of formation of these nodules are greater than l600°C. This is greater than any lherzolite solidus at pressures where spinel is stable, even under anhydrous conditions (Kushiro et. al. 1968; Nishikawa et. al. 1970). While there is some doubt as to the absolute temperatures at which the nodules formed, the relative temperatures are believed to be correct (p.60) so that these nodules appear to befrefractory and may be residue from partial fusion of the mantle. Textural evidence described in Chapter 3 suggests that the above 20001 o IOOO • o_ e a> 500 IO 20 30 40 50 Pressure in Kb. Various lherzolite solidi. A •. Anhydrous \ B : PH2O < piotai ( Kushiro I968 ) C : PH20 = Ptotal J D: Anhydrous ( Nishikawa I970 ) O Fig. 3 3 : .91 discussion is reasonable. It has been suggested that the marginal alteration of the diopsides of the Castle Rock and Jacques Lake nodules (Fig. 8) is due to the reactions jadeitlc diopside —> jadeite-poor diopside + feldspar If the reaction is of this form, then another phase must have participated (either as a reactant or a catalyst) since the reaction occurs only at the rims of the diopsides. Enclosing minerals such as olivine are unaffected (Fig. 10). A fluid phase could have been present, but if a fluid was present and participated in the reaction, then coexisting orthopyroxene should also have > been affected since it has been shown that orthopyroxene is unstable at thermargin of the nodules. Coexisting olivine and enstatite are unaffected by any reaction (Fig,10), Therefore the breakdown of diopside is not due to reaction with a fluid phase and cannot be a polymorphic change due to variation In the pressure and temperature since reaction occurs only at the rims of the diopsides. One explanation of this is that it is due to partial melting. It is possible that the liquid produced by partial melting of lherzolite was trapped and quenched when the nodules were brought to the surface. The glass then devitrifled to form feldspar (in part at least). Diopside is the first phase to melt in a lherzolite of likely mantle composition (Kushiro 1969b; Ito and Kennedy 1967). Dickey et. al. (1971) have found that Cr-bearing diopside melts incongr-uently to spinel and liquid above 5Kb. The effects of A1203, Cr20<j and other components on melting behavior at high pressure in the system Di-Fo-S102 (a simplified peridotite system) which .92 was studied by Kushiro (1969b) are not known but in view of the behavior of Cr-bearing diopside (Dickey et. al. 1971) they are likely to be significant. It is not necessary for the minimum melting point of lherzolite to be a eutectic. In the simplified peridotite system it is hot certain whether point P (Fig.3*0 is a eutectic or a reaction point.cEdiptiAtOB* the Di'-Fo join is a piercing point (Fig.34) (Kushiro 1969b). Thus melting of diopside in a natural system is a possible explanation of the described texture, • • . The weight of the evidence favours the hypothesis that the source of the Castle Rock and Jacques Lake nodules is the upper mantle and that these nodules could be fragments of the mantle which has been depleted by partial melting. There are two lines of evidence to suggest that the source of the Castle Rock nodules is layered. The majority of the samples available for study show some mineraloglcal layering. The" olivine fabric of one of the layered specimens appears to be related to the layering (Fig.20). The massive specimens have a similar fabric which suggests that the mechanism which produced the layering was operative throughout the source rock, even though some of hand specimen sized nodules are apparently unlayered. Mesoscopic layering is therefore probably characteristic of the source of the Castle Rock nodules. On the other hand, the source rock of the Jacques Lake nodules is unlikely to be layered, at least on a mesoscopic scale. Only a few of several hundred specimens observed in the field were layered. The olivine fabric of the nodules which were studied is different from the Castle Rock fabrics or other layered ultramaflc rocks (Chapter 4). CaMgSi206 CaMgSi206 Fig. 34 : Part of the liquidus diagram of the system Fo-Di-SiC>2 at 20Kb. pressure j (a) hydrous , (b) anhydrous. (after Kushiro 1969b). .94 CHAPTER 9 The Upper Mantle In British Columbia. It has been shown that the lherzolite nodules of Castle Rock and Jacques Lake are probably samples of the upper mantle. The differences and similarities between these suites of rocks are significant with respect to the constitution of the upper mantle in British Columbia. While much of the following discussion is speculative, it is nevertheless useful as a guide to what might be expected upon further study of the mantle in British Columbia. The limited evidence from this study suggests that the upper mantle in British Columbia consists largely of spinel lherzolite. There is no evidence of regional differences in mineralogy nor of mineralogical zoning. There is evidence of chemical variations. The Castle Rock pyroxenes are more magnesian and the spinels and pyroxenes are more aluminous than the corresponding minerals of the Jacques Lake nodules. This may be a result of a fundamental difference in the chemistry of the mantle in these.- areas or may be a result of different degrees of partial melting. The Castle Rock nodules are more refractory so if the upper mantle in British Columbia was originally homogeneous,different degrees of partial melting have resulted in the differences in mineral chemistry. As well as major element variations there are variations in the trace element concentrations in the two areas. Both Castle Rock pyroxenes contain more Co^p'theediopsides contain more Ni and Zn, and the enstatites more Mn than the corresponing minerals of the Jacques Lake nodules. The upper mantle beneath Castle Rock .95 appears to be enriched in these elements relative to the mantle beneath Jacques Lake. This enrichment Is probably a primary feature of the mantle in the Castle Rock region since the concentration of these elements is independent of major element variations. Further sampling of nodules and also of the basaltic rocks of each area is required to support this suggestion. Analyses for more mobile elements such as Rb, Sr and the rare earths would be useful to test the hypothesis'; that geochenical provinces exist in the upper mantle in British Columbia.- Data on these elements might allow one to evaluate the extent of partial melting and so distinguish primary variations in the chemistry of the mantle from variations due to different degrees of partial melting. Physical as well as chemical variations exist in the upper mantle in'.British Columbia. It has been shown that the olivine : fabrics of the Castle Rock and Jacques Lake nodules have resulted from deformation in the solid state, and that the suites have ->r* different characteristic fabrics. The fabrics are considered to have been imposed on the rocks prior to their Inclusion in their enclosing rocks and are a result of stress within the mantle. Hess (1964) showed that seismic anisotropy in the upper oceanic mantle is caused by the alignment of olivine in the direction of flow at major fracture zones. Keen and Barrett (1971) found that the mantle west of the Queen Charlotte Islands has an anisotropy related to the direction of sea floor spreading (Fig. 35). It is possible that this anisotropy extends into the mantle beneath British Columbia. Both the Castle Rock and Jacques Lake .96 olivines have a strong degree of prefered orientation suggesting that both regions are underlain by an anisotropic mantle. Each suite has a different fabric which suggests that the degeee of anisotropy is not the same in each area. Also,the orientation of the olivine fabrics in the mantle and consequently the seismic velocity vectors in each area are likely to be different. Souther (1970) has suggested that the regional distribution of Quaternary volcanoes in British Columbia is controlled by major faults associated with mantle structures>, (Pig. 3*0 • If the orientation of olivine in the mantle is associated with fracture zones in the mantle, then different stress regimes associated with the two regions will produce different fabrics in the olivines and will result in varyingltdegrees of anisotropy. Castle Rock appears to lie in a zone of extension and Jacques Lake appears to lie in a zone of shear above the mantle. (Fig.35). This might result in different fabrics being produced in the olivines in the mantle in these areas. Other factors such as strain rate, temperature, pressure and the presence or absence of water will also contribute to the type of fabric developed by a particular suite of rocks. Which of these is dominant is not known. Whatever the cause, it is likely that variations in the degree of anisotropy and consequently in the seismic behavior of the mantle in British Columbia are to be expected and that these may eventually be integrated into the regional tectonic framework. Fig. 3 5 Major structural features related to recent volcanism in British Columbia, (after Souther 1970) Belts of Quaternary volcanoes, j Eastern limit of Tertiary and Recent transcurrent ' faulting (right lateral shear). Direction of inferred relative motion between adjacent segments of the Cordillera. A Area of mantle anisotropy. ( Keen and Barret 1971) .98 CHAPTER 10• Conclusions. The main finding of this study is that each suite of lherzolite nodules in basaltic rocks from British Columbia is characterised by its own range of mineral compositions and fabrics. The range of compositions for each suite is narrow and overlap to some extent. The distribution of some elements (eg. Fe, Ni,Mn,Co,Zn) between the minerals of each suite is different and is independent of mineral composition. The significance of this is that each suite probably formed under different P/T conditions. Comparison with other studies and with relevant experimental work places the source of these nodules in the upper mantle. This agrees with the P/T conditions inferred from the mineral chemistry of these nodules. Consideration of the textures and olivine fabrics of the Castle Rock nodules suggests that these rocks are fragments of the refractory upper mantle which is layered and has been deformed. The Jacques Lake nodules are also residual fragments of the mantle which has been deformed but in the Jacques Lake area Is uhlayered. The mineral chemistry of the nodules has shown that the Castle Rock lherzolites formed at higher temperatures and probably at greater depths than those of Jacques Lake. The Nicola Lake nodules are crystal accumulates, have not been deformed, and formed at lower pressures and temperatures than the other nodules, They could be cognate with their present host rocks. The different chemical characteristics of the Castle Rock and .99 Jacques Lake nodules suggest that the mantle is chemically different in these regions. This may be a result of different degrees of partial melting of an originally homogeneous mantle but may also be a result of original heterogeneity The strong prefered orientation of the olivine of the Castle Rock lherzolites and, to a lesser extent, those of Jacques Lake suggests that the mantle in these two regions is anisotropic. The different fabric types suggest that different deformational regimes are to be found from place to place in the mantle. .100 BIBLIOGRAPHY Aokl,K. (1968). Petrogenesls of ultrabasic and basic inclusions in alkali basalts, Iki Island, Japan. Am. Miner. 53, 241-256. Aoki.K. and Kushiro,I. (1968). Some clinopyroxenes from ultra mafic Inclusions In Dreiser Weiher, Elfel. Contr. Miner. Petrol. 18, 326-337. Ave'Lallement,H. and Carter,N.L. (1970). Syntectonic recrystall-isation of olivine and modes of flow in the upper mantle. Bull. geol. Soc. Am. 81, 2203-2220. Bartholome.P. (1962). Iron-magnesium ratio in associated pyroxenes and olivines. In Petrologic Studies; Geol. Soc. Am. Buddington Volume, 1-20. Binns,R.A. (1969). High pressure megacrysts in basanitic lavas near Armidale, New South Wales. Am. J. Sci. Schairer Volume 267-A. 33-49. Black,P.M. and Brothers,R,N. (1965). Olivine nodules in olivine nephelinite from Tokatoka, Northland. New Zealand J. Geol. Geophys. 8, 62-80. Boyd.F.R. (1963). Some effects of pressure on phase relation ships in the system Mg0-Al203-Si02. Yb. Carnegie Inst. Wash. 62, 121-124. Boyd,F.R. and England,J.L. (i960). Aluminous enstatite. Yb. Carnegie Inst. Wash. J5£, 49-52. Brothers,R.N. (1959). Flow orientation of olivine. Am. J. Sci. 2 57. 574-584. Brothers,R.N. (i960). Olivine nodules from New Zealand. Proc. 21st. Int. geol. Congr. lji, 68-81. Brothers,R.N. (1962). The relationship between prefered orient ation of olivine in dunlte and the tectonic environments A discussion. Am. J. Sci. 260, 310-312. Brothers,R.N. (1964). Petrofabric analyses of Rhum and Skaergaard layered rocks. J. Petrology j>, 255-274. Brothers,R.N. and Rodgers.K.A. (1969). Petrofabric studies of ultramafic nodules from Aukland, New Zealand. J. Geol. 22, 452-465. Burns,R.G. (1970). Mlneralogical applications of crystal field theory. London» Cambridge University Press. .101 Cambell.R.B. (196l). Quesnel Lake, West Half, Cariboo District, British Columbia. Geol. Surv..Canada,Map 3-1961. Carter,J.L. (1970). Mineralogy and chemistry of the earth's upper mantle, based on the partial fusion - partial crystallisation model. Bull. geol. Soc. Am. 81^, 2021-20^4. , Carter,N.L. and Ave»Lallement,H. (1970). High temperature flow of dunlte and peridotite. Bull. geol. Soc. Am. 8l, 2181-2202. Collee.A.I.G. (1963). Ailfabric study of lherzolites with special reference to nodular inclusions in the lavas of Auvergne, France. Leld. geol. Meded. 28, 1-102. Dickey,J.S., Yoder.H.S. and Schairer,J,F. (1971). Chromium in silicate-oxide systems. Yb. Carnegie Inst. Wash. 70» 118-125. Forbes,R.B. and Banno.S. (1966). Nickel-iron content of peridotite inclusions and cognate olivine from an alkali-olivine basalt. Am. Miner. 41, 130-140. Forbes,R.B. and Kuno.H. (1965). The regional petrology of peridotite inclusions and basaltic host rocks. Upper mantle symposium, New Delhi 1964. Copenhagen: Int. Un. geol. Scl. , 161-179. Forbes,R.B. and Kuno.H. (1967). Peridotite inclusions and basaltic host rocks. In Ultramaflc and Related Rocks, 328-337. Wyllie.P.J. (ed.). New York: Wiley. Fuster,J.M.,Paez,A. and Sagredo,J. (1969). Significance of basic and ultramaflc Inclusions in the basalts of the Canary Islands. Bull. Volcan. 22» 665-693. Goles,G.C. (1967). Trace elements in ultramaflc rocks. In Ultra-mafic and Related Rocks, 352-362. Wyllie,P.J. (ed.). New York: Wiley. Grover.G.E. and Orville.P.M. (1969). The partitioning of cations between coexisting single and phases with applications to the assemblages: orthopyroxene- clino pyroxene and orthopyroxene-olivine. Geochim. cosmochim. Acta 21* 205-226. Green,D.H. (1968). Origin of basalt magmas. In Basalts Vol.2, 835-862. Hess.H. and Poldervaart,A. (eds.). New York: Interscience. .102 Green,D.H, and Ringwood,A.E. (1969). The origin of basaltic magmas. Monogr. Am. geophys. Un. 1^, 489-495. Green,D.H. and Ringwood,A.E. (1970). Mineralogy of peridotltic compositions under upper mantle conditions. Phys. Earth planet. Interiors 359-371. Hamad,El.D. (1963). The chemistry and mineralogy of the olivine nodules of Colton Hill, Derbyshire. Mineralog. Mag. 3"), 48.3-497. Harris,P.G.,Reay,A. and White,I.G. (1967). Chemical composition of the upper mantle. J. geophys. Res. 7_2» 6359-6369. Hess.H.H. #1964). Seismic anisotropy of the uppermost mantle under oceans. Nature 203> 629-631. Irvine,T.N. (1965). Chromian spinel as a petrogenetlc indicator. Part Is Theory. Can. J. Earth Sci. 2, 648-672. Irvine,T.N. (1967). Chromian spinel as a petrogenetlc indicator. Part 2s Petrologic applications. Can. J. Earth Sci. 4, 71-103. Ishibashi.K. (1970). Petrochemical study of basic and ultrabaslc Inclusions In basaltic rocks from Northern Kyushu, Japan. Mem. Fac. Sci., Kyushu Univ., Ser.D.Vol. 20. 85-146. Ito.K. and Kennedy,G.C. (1967). Melting of peridotite at 40Kb. Am. J. Sci. 26j>, 519-538. Jackson,E.D. (196l). Primary textures and mineral associations in the ultramaflc zonetiof the Stillwater Complex, Montana. Prof. Pap. U.S. geol. Surv. 358. Jackson,E.D. (1968). The character of the lower crust and upper mantle beneath the Hawaiian Islands. Proc. 23rd. Int. geol. Congr. _1, 135-150. Jackson,E.D. (1969). Chemical variation in coexisting chromlte and olivine in the chromite zones of the Stillwater Complex. Symposium on Magmatic Ore Deposits, Econ. Geol. Monogr. 4, 41-71. Jackson,E.D. and Wright,T.L. (1970). Xenollths in the Honolulu Volcanic Series, Hawaii. J. Petrology U, 405-430. Keen,C.E. and Barrett.D.L. (1971). A measurement of seismic anisotropy in the Northeast Pacific.! Can. J. Earth Sci. 8, 1056-1064. Kretz,R. (1961). Some applications of thermodynamics to coexist ing minerals of variable composition. Examples $ ortho-pyroxene-clinopy-roxene and orthopyroxene-garnet. J. Geol. 69, 361-387. .103 Kretz,R. (1963). Distribution of magnesium and iron between orthopyroxene and calcic pyroxene in natural mineral assemblages. J. Geol. 7_1, 773-785. Kuno,H. (1967). Mafic and ultramafic nodules from Itonome-gata, Japan. In Ultramafic and Related Rocks, 337-342. Wyllle, P.J. (ed.). New York: Wiley. Kuno.H. (1969). Mafic and ultramafic nodules in basaltic rocks of Hawaii. Mem. geol. Soc. Am. 11J>, 189-234. Kuno,H. and Aoki.K. (1970). Chemistry of ultramafic nodules and their bearing on the origin of basaltic magmas. Phys. Earth planet. Interiors 273-301. Kushiro,I. (1969a). Clinopyroxene solid solutions formed by reactions between diopside and plagioclase at high pressures. Spec. Pap. mineralog. Soc. Am. 2, 179-191. Kushiro,I. (1969b). The system forsterite-dlopside-sillca with and without water at high pressures. Am. J. Sci. Schalrer Vol. 267-A, 269-294. Kushiro,I. and Yoder.H.S. (1966). Anorthite-forsterite and anorthite-enstatite reactions and their bearing on the basalt -eclogite transformation. J. Petrology 7_, 337-362. Kushiro,1..Yasuhikb,S. and Akimoto.S. (1968). Melting of a peridotite at high pressures and high water pressures. J. geophys. Res. 22* 6023-6029. Kutolin.V.A. and Frolova.V.M. (1970). Petrology of ultramafic inclusions from basalts of Minusa and Transbaikallan Regions. Contr. Miner. Petrol. 2£, 163-179. Langmyhr,F.J. and Paus,P.E. (1968). The analysis of inorganic siliceous materials by atomic absorption spectrophotometry and the hydrofluoric acid decomposition technique. Part 1: The analysis of silicate rocks. Analytlca chimlca Acta 4j, 397-408. Loney,R.A.,Himmelberg,G.R. and Coleman,R.G. (1971). Structure and petrology of the Alpine-type peridotite at Burro Mountain, California, U.S.A. J. Petrology 12, 245-309. Matsui.Y, and Banno.S. (1970). Partition of divalent transition metals between coexisting ferromagnesian minerals. Chem. Geol. j>, 259-265. Mclntyre,W.I. (1963). Trace element partition coefficients. A review of theory and applications to geology. Geochlm. cosmochim. Acta 27_, 1209-1264. .104 McGregor,I.D. (1967). Mineralogy of model mantle compositions. In Ultramafic and Related Rocks, 382-393. Wyllie.P.J. [cy (ed.). New Yorki Wiley. McGregor,I.D. (1968). Mafic and ultramafic inclusions as indicators of the depth of origin of basaltic magmas. J. geophys. Res. 22. 3737-3745. McGregor,I.D. (1970). The effect of CaO, Cr20o, Fe20o and AI0O3 on the stability of spinel and garnet peridotites. Phys. Earth planet. Interiors 3, 372-377. Medaris,L.G. (1969). Partitioning of Fe2+ and Mg2+ between coexisting synthetic olivine and orthopyroxene. Am. J. Sci. 267, 945-968. Mercy,E. and O'Hara.M.J. (1967). Distribution of Mn, Cr, Ti and Ni in coexisting minerals of ultramafic rocks. Geochim. cosmochim. Acta 2331-2341. Mueller,R.F. (196l). Analysis of relations among Mg, Fe and Mn in certain metamorphic minerals. Geochim. cosmochim. Acta 2£, 267-296. Nafziger,R.H. and Muan,A. (1967). Equilibrium phase compositions and thermodynamic properties of olivines and pyroxenes in the system Mg0-"Fe0"-Si02. Am. Miner. ^2_, 1364-1385. Nlshlkawa.M.,Kono,S. and Aramaki.S. (1970). Melting of lherzolite from Itonome-gata at high pressures. Phys. Earth planet. Interiors 4, 138-144. O'Hara.M.J. (1963). Distribution of iron between olivines and calcium-poor pyroxenes in peridotites, gabbros and other magnesian environments. Am. J. Sci. 26l, 32-46. O'Hara.M.J. (1967). Crystal-liquid equilibria and the origins of ultramafic nodules in basic Igneous rocks. In Ultra mafic and Related Rocks, 346-349. Wyllie.P.J. (ed.). New York: Wiley. O'Hara.M.J. (1968). The bearing of phase equilibria studies in synthetic and natural systems on the origins of basic and ultrabasic rocks. Earth Sci. Rev. 4£ 69-133. Presnall,D.C. (1969). The geometrical analysis of partial fusion. Am. J. Sci. 262, 1178-1194. Ragan.D.M. (1969). Olivine recrystallisation textures. Mineralog. Mag. 3Z, 238-240. Raleigh,C.B. (1968). Mechanism of plastic deformation of olivine. J. geophys. Res. 21* 5391-5406. 0 .105 Ram"berg,H. and DeVore.G.W. (195D. The distribution of Fe2 + and Mg in coexisting olivines and pyroxenes. J. Geol. J§2. 193-210. Ringwood,A.E. (1966). Mineralogy of the mantle. In Advances in Earth Science, 357-399. Hurley,P. (ed.). Cambridge: M.I.T. Press. Ringwood,A.E. (1969). Composition and evolution of the upper mantle. Monogr. Am. geophys. Un. 1^, 1-17. Roedder.E. (1965). Liquid CO2 inclusions in olivine-bearing nodules from basalts. Am. Miner. j>0, 1746-1782. Ross.C.S., Foster,M.D. and Myers,A.T. (1954). Origin of dunites and of olivine-rich inclusions In basaltic rocks. Am. Miner. 2£, 693-736. Rucklidge,J. and Gasparrini,E.L. (1969). Specifications of a computer program for processing electron microprobe analytical data. EMPADR Vll. Dept. of Geol., University of Toronto, Toronto, Ontario. Sandell,E.B. (1959). Colorimetric Determinations of Traces of Metals. New York: Interscience. Snedecor,G.W. and Covhran,W.G. (1967). Statistical Methods. Ames: Iowa State University Press. Simkin.T. and Smith,J.V. (1970). Minor- element distribution in olivine. J. Geol. 7jB, 304-325. Soregaroll,A. (1968). Geology of Boss Mountain Mine, British Columbia. Ph.D. thesis. University of British Columbia, Vancouver. Souther,J.G. (1970). Volcanism and its relationship to recent crustal movements in the Canadian Cordillera. Can. J. Earth Sci. 2, 553-568. Talbot,J.L, Hobbs,B.E., Wilshire,H.G. and Sweatman,T.R. (1963). Xenollths and xenocrysts from lavas of the Kerguelen Island Archlpeligo. Am. Miner. 48, 159-179. Thayer,T.P. (i960). Some critical differences between Alpine-type and stratiform peridotite-gabbro complexes. Proc. 21st. Int. geolJ. Congr. 1^, 247-259. Tredger.P. (1970). Petrology of nodules in olivine basalt from Quesnel Lake, British Columbia. . B.A.Sc. thesis. University of British Columbia, Vancouver. .106 Turner,P.J. (1942). Prefered orientation of olivine crystals in peridotites with special reference to New Zealand examples. Trans, and Proc. Roy. Soc. New Zealand 72, 280-300. Wllshire.H.G. and Binns.R.A. (196l). Basic and ultrabasic xenollths from volcanic rocks of New South Wales. J. .-eifel-* Petrology 2, 185-208. White,R.W. (1966). Ultramaflc inclusions in basaltic rocks from Hawaii. Contr. Miner. Petrol. 12, 245-314. YamaguchijM. (1964). Petrogenetlc significance of ultrabasic inclusions in basaltic rocks from Southwest Japan. Mem. Fac. Sci., Kyushu Univ., Ser.D. Vol.20. 163-219. .107 APPENDIX 1 Analytical Techniques. (a) Mineral separation and sample preparation. The constituent minerals of each nodule were separated by a combination of hand sorting and by use of a Franz magnetic separator. Purity was estimated by point-counting grains on a 1mm. transparent grid. The final purity of the mineral separates was greater than 99.5$ in most cases and never less than 99.0$. About lg portions of olivine, orthopyroxene and clinopyroxene were ground to a fine powder (-100 to -200 mesh) by hand for 15 minutes in an agate mortar. A few grains of each spinel were mounted in Fibrolay epoxy and polished with tin oxide. Prior to the probe analyses each mount was coated withaa thin layer of carbon. (b) Electron microprobe analyses. The s-piireflXanalyses were carried out with a JXA-3 electron microprobe X-ray analyser. The analyses were done by comparing Intensities (measured as counts per second) of selected X-ray lines to those from standards of known composition. In all cases first order Koc lines were used. The voltage was 25Kv. for every element. A 10 second counting time was used in each case,10 to '20 points on each grain being analysed. The average of each series of counts was taken as the true intensity. The standards were analysed before and after each run to determine instrumental drift. After each run the background was determined for both standards and samples. Table 1 lists, the elements which were determined, refered to the appropriate standard. .108 TABLE 1 Elements and standards used In electron microprobe analyses. Element Standard Pe Pure Fe metal Cr Pure Cr metal Mg Synthetic spinel* Al - " * Composition is : MgO 28.30; AlgO^ 71.55; FeO 0.02; CaO 0.02. TABLE 2 Operating conditions for the hollow cathode lamps. Element Wavelength^) Lamp current(ma.) Slit( ) Flame Co 2407 5 25 acetylene-air Cu 3247 3 50 Mn 2794 10 100 Ni 2320 8 50 Pb 2170 6 300 Zn 2138 6 100 Na* 5890 5 200 K* 7664 10 200 Ca# 4226 10 25 •• Mg# 2852 4 50 " Fe 3719 5 50 Ti 3643 20 100 acetylene-nitrous oxide Al 3091 11 100. 11 *Cs added to samples and standards to suppress Interferences. #La added to samples and standards to suppress interferences. The data were processed through the computer program EMPADR VI1 which applies corrections for background, dead time if necessary, atomic number, absorption and fluorescence and converts the readings to weight percent of the appropriate oxide. (Rucklidge and Gasparrini 1969). The precision of the analyses for each element calculated as .109 the standard error of the mean of each series of counts, is given below. For Al and Mg the error is close to $% of the amount present and for Fe and Cr it is about 1%. The error is consistent from sample to sample except for Al which varies from k to 6%, Accuracy can be no better than precision so that the errors in counting alone can account for the deviations from 100$ in the totals. (c) Atomic-absorption analyses. All elements expected to have a concentration of less than 2% were treated as trace elements in the analytical scheme. These were Co,Cu,Mn,Ni,Pb,Zn,Na,K,Ti,and Ca (in olivine). 0.3000g of the mineral powder was dissolved in 5ml of HF and 1ml of HCIO^ and the solution evaporated to dryness at 180°C on a hotplate. The residue was taken up in 3ml °f HC1 and the solution made up to 25ml with distilled water. Series of standards of appropriate concentration were made up in 1.5M HC1. The standards were aspirated into the flame of a Techtron AA-4 atomic-absorption spectrophotometer and a plot of concentration versus absorption prepared. The samples were then aspirated, after appropriate dilution if necessary, and the concentration read from the graph. The operating conditions of the hollow cathode lamps are summarised in Table 2. For major elements (Mg,Fe,Al and Ca in pyroxenes) a method described by Langmyhr and Paus (1968) was used. 0.2000g of mineral powder was dissolved in 5ml of HF and evaporated to dryness at 180°C on a hotplate. A further 5ml of HF was then added and the solution warmed. 50ml of saturated boric acid solution was then added to dissolve the precipitated fluorides and to complex any .110 excess HF. The solution was made up to 100ml with distilled water. Standards of appropriate concentration were made up in the same way. Each sample solution was aspirated four times into the flame of a Techtron AA-4 atomic-absorption spectrophotometer, bracketing It each time between standards of appropriate concentration. The order of aspiration was reversed after each set of readings. The concentration of the elements in each sample was calculated from the following equations Ex - El C = A + K E2 _ EI where A is the concentration (wt.$) in the lower standard, K the difference in weight between the upper and lower standards, Ex, El and E2 are the absorbance of the sample solution, lower standard and upper standard respectively. The arithmetic mean of the four readings was taken as the concentration. Operating conditions for the hollow cathode lamps are summarised in Table 2. Each batch of samples included a duplicate and a blank. No corrections for the blank were necessary. No corrections for background were required for any of the elements. The precision of all the analyses was estimated as the standard deviation of the duplicate analyses. For all elements the preci sion was better than $%% and generally about 3% of the amount present. The largest errors were in Mg due. to the high dilution factor required and the sensitity of the lamp, and in Al which Is sensitive to the fuel flow, (d) Determination of C^O^. Chromium was determined colorometrically using an adaption of .111 the methods described in Sandell (1967). A 20ml aliquot of the solution used for the determination of Mg etc. by atomic-absorption was taken. 5ml of 6N H2SOij, was added and the solution warmed. A few drops of 0.1N KMnO^ solution was added to this until the solution remained faintly pink on heating. The solution was boiled for ten minutes, allowed to cool and O.lg of Na20 and 10 - 20ml of 20$ NaCO^ solution added until a permanent precipitate appeared. The solution was then boiled for ten minutes, cooled and filtered. Enough 6N H^JSO^ (10 -20ml) was added carefully with swirling to liberate C02 until the _ TvX solution was approximately 0.2N in H2S04. 1ml of diphenylcarbazide solution was added and the solution made up to 50ml with distilled water. The purple colour so obtained was compared visually to a series of standardsisolutions containing 0.2 - 5ppm Cr made up with standard K2Cr20y in the same way as the sample solutions. Duplicate samples and a blank solution were run with each batch. Precision, determined as the standard deviation of the average of the duplicates, is 5$ of the amount present. Due to the dilute solutions used, the limit of detectability Is 2'60ppm. Consequently the chromium content of the olivines was not determined. A slight error is Introduced by using this method as some Cr is driven off as a fluoride during the initial decomposition. However this is not believed to be serious as the pyroxenes contain between 0.5 and 1.0$ Cr20^ and reproducibility is only 5$. Precision could be improved by making up fresh concentrated solu tions, but for rapid convenient analyses, the above method is considered adequate in view of the errors- in the other elements, .112 APPENDIX 2 Error propagation In temperature determination. The precision in determining Mg and Pe in olivine (based on duplicate analyses) is less than 5% of the amount present. The precision in determining Mg and Al in the spinel is 5% of the amount present; for Fe and Cr precision is 1% of the amount present. These figures are based on counting statistics. Table 1 gives the uncertainties in ratios involving these elements, based on the above precisions. Table 1 Uncertainties in element ratios. Ratio Uncertainty Ratio Uncertainty ol * .004 Cr Cr + Al + Fe Al Cr + Al * Fe Fe3+ Cr + Al + Fe Mg + Fe Mg Mg + Fe2+ SP - 'OOS Fe2+ Mg + Fe2+ sp - .005 Cr ~ * Uncertainty taken as the uncertainty in Cr + Al + FeJ as Fe20<j calculated by assuming the model spinel formula. The uncertainties in KD(4) were calculated from the expression: |dy | /fx1(x1... .xn)/ / dxx/ + /fx2(x1... .x^)// dx2| + /fxn(x1... .xn)/ /dxn/ where dy was set equal to dKD(4), x^ to X°^, x2 to Xp^, x^ to Xpg and x^ to X^.. dx^, dx2, dx^and dx^ were taken from Table 1. (these were calculated using the above equation with the appropriate substitutions). The uncertainties in the derived temperatures were calculated from the above equation where dy was set.equal to dT, x^ to x^ to the variables on the right hand side of the equation for the calculation of the temperatures _ 5580*+ 1018/9- 1720T + 2400  .90* + 2,56/3 - 3.08^- 1.47 + 1.98?lnKD(4) and dx1 to dx7 were taken from Table 1,, and Table 7 in the text. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics

Country Views Downloads
Germany 21 16
China 21 31
United States 9 0
Russia 8 0
Canada 5 1
Ukraine 2 0
Japan 2 0
City Views Downloads
Unknown 23 16
Beijing 13 0
Saint Petersburg 8 0
Ashburn 8 0
Shenzhen 8 31
Montreal 4 0
Tokyo 2 0
Spokane 1 0
Greater Sudbury 1 1

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items