28. CHEMISTRY OF ALTERATION MINERALS FROM DEEP SEA DRILLING PROJECTSITES 501, 504, AND 5051

A. C. Adamson, Department of Geology, The University, Newcastle upon Tyne, NE1 7RU, England

ABSTRACT

Over 300 electron microprobe and wet chemical analyses were made of alteration minerals (mostly clays) in basaltsfrom DSDP Sites 501, 504, and 505 in the Costa Rica Rift. These analyses are displayed using a three-dimensionalgraphical method. Several alteration minerals have been identified, of which saponites are by far the most common. Insitu alteration of basalt glass has produced montmorillonite-like minerals rich in the less-mobile elements such as Al andTi. Talc-like minerals commonly replace olivine and act as fillings to vesicles. Three other alteration minerals have beenrecognized chemically, rich in Na and Ca, but these are presently unidentified. There is no evidence for systematicchange of mineral chemistry with depth.

INTRODUCTION

The alteration of ocean-floor basalts by sea water hasbeen studied in considerable detail in rocks from manydifferent environments since the beginning of the DeepSea Drilling Project (Andrews, 1977; Baragar, et al.,1977; Bass, 1976; Bohlke, et al., 1980; Donnelly, 1978;Hart, 1973; Peterson, 1978; Pritchard, 1979; Pritchardet al., 1978; Scarfe and Smith, 1977; Scott and Hajash,1976). None of these, however, have specifically at-tempted to distinguish between or investigate differ-ences in the alteration chemistry of sites with contrast-ing geothermal environments. DSDP Legs 68, 69, and70 made the first positive attempt to remedy this situa-tion, by studying in detail a section of the ocean floor onthe southern flank of the Costa Rica Rift, where both"open" and "closed" hydrothermal circulation wasthought to be operating. For the first time, these legshave allowed the opportunity to study basalt alterationin rocks of similar petrographic type, but from two op-posite geothermal environments.

This study is essentially a geochemical one; the petro-graphic relationships between the various alterationphases are described elsewhere (Honnorez et al.; Noacket al.; Kurnosov et al.; Pertsev and Boronikhin, all thisvolume). Changes in alteration chemistry are identifiedon a local and extended scale, with the aim of recog-nizing relationships which correspond to an increase indepth or a change in the local petrographic environ-ment. The fine layering apparent in many precipitatedclays has been analyzed in detail, and the geochemicalvariations have been used as an indicator of local fluc-tuations in fluid chemistry with time.

Samples were selected from areas of the core whereveins, oxidation rims, pseudomorphs, breccias, and cav-ity and vesicle fillings were particularly abundant, indi-

Cann, J. R., Langseth, M. J., Honnorez, J., Von Herzen, R. D., White, S. M., et al.,Init. Repts. DSDP, 69: Washington (U.S. Govt. Printing Office).

eating intense alteration. The resulting bias of this sam-pling should be remembered when reading this report.

Over 300 electron microprobe analyses were collectedof various alteration phases from 23 thin-sections evenlydistributed in cores from Holes 504A, 504B, 505, and505B. A further 25 analyses were made using wet chemi-cal methods on separates from samples rich in clay. Fer-rous iron content was determined by normal colorime-tric means. Electron-microprobe analysis was conductedon microprobe equipment with energy-dispersive spec-trometry (Department of Earth Sciences, Cambridge Un-iversity, England), producing 14 element analyses. X-ray diffractograms and powder photographs made witha 114.6-mm Debye-Scherrer camera and CoKα radia-tion with Fe filter have aided in the identification andgrouping of the various alteration minerals.

A three-dimensional graphical method has been de-veloped to enable rapid categorization of large numbersof unknown-clay analyses. Alteration products are con-sidered to be 2:1 layered clays, and analyses are recalcu-lated on the basis of 22 oxygens with the inter-layercharge and total Al + Fe3 + determined. Depicted graph-ically, similar analyses form clusters. The introductionof a third axis, measuring the dioctahedral, trioctahed-ral, or intermediate character of analyses, subdividesclusters into linear trends along which members showsystematic changes in substitution and chemistry.

The method has allowed several clay types to beidentified, most notably saponite, montmorillonite-likeclays, talc-like clays, and various mixed layer phases.Three other alteration minerals, rich in Na and Ca, havebeen recognized chemically but remain unidentified.

In general, the findings of this study are in line withthose of previous work on low-temperature alteration ofbasalts, such as at Holes 417A and 418A, Legs 51, 52,and 53, as described by Pritchard (1979); Pertsev andRusinov (1979), Rusinov et al. (1979), Donnelly, Prit-chard, et al. (1979), and Donnelly, Thompson, et al.(1979).

551

A. C. ADAMSON

A THREE-DIMENSIONAL GRAPHICAL METHODFOR THE CATEGORIZATION OF

CLAY ANALYSES

The structures of hydrous layered silicates, commonlyknown as clay minerals, are based upon continuous two-dimensional sheets of SiO4 tetrahedra, individual tetra-hedra linked with neighbors by sharing the three basalcorners in six-fold symmetry (Bailey, 1980). The fourthtetrahedral corners (apical oxygens) all point in the samedirection normal to the sheet and at the same time formpart of an immediately adjacent octahedral layer inwhich individual octahedra are linked laterally by shar-ing octahedral edges. The common plane of junction be-tween the tetrahedral and octahedral sheets consists ofthe shared apical oxygens plus unshared OH groups thatlie at the center of each tetrahedral six-fold ring. The oc-tahedral cations are normally Mg, Al, Fe2+, and Fe3 + ,but other medium-sized cations may occur in some spe-cies. The smallest structural unit contains three octa-hedra occupied by divalent or trivalent elements. Acharge balance is achieved either when all three octa-hedra are occupied by divalent octahedral cations (thesheet is then classified as trioctahedral), or when two ofthe tetrahedra are occupied by trivalent ions and thethird is vacant (the sheet is classified dioctahedral).

The layered silicates are built up of different combin-ations of these two types of sheet. A 2:1 layer links twotetrahedral sheets with one octahedral sheet. This re-quires the upper tetrahedral sheet to be inverted, so thatits apical oxygens point down and can be shared with theoctahedral sheet below. If the 2:1 layer is not electro-statically neutral, the excess layer charge may be neu-tralized by inter-layer materials, including individual ca-tions, hydrated cations, and hydroxide octahedral groupsand sheets.

The chemical composition of 2:1 layered clay variesaccording to the extent of cation substitution in bothtetrahedral and octahedral sites, and the nature andquantity of inter-layer materials. Substitution of ions ofthe same valence, notably Mg-Fe2+ and Al-Fe3+ substi-tutions, are common in octahedral positions. Substitu-tions by ions of lesser charge, notably Si4+ by Al3+ intetrahedral positions, and Al3+ or Fe3+ by Mg or Fe2+

in octahedral positions, produce negative charges on thelayers which are balanced by inter-layer materials.

The principal 2:1 layered clays with their extremes ofcharge per formula unit (x) and types of inter-layer ca-tions are:

1) Illite group, x = 1.0 to 1.5, K+ as the principalinter-layer ion;

2) Vermiculite group, x = 0.7 to 1.0, Mg2+ and Ca2+

as the principal inter-layer ions;3) Smectite group, x = 0.2 to 0.7, Ca2+ and Na+ as

the principal inter-layer ions.The structure of 2:1 layered clays does not allow in-

ter-layer charge to exceed 2.0, and a further constrainton composition comes from the observation that totalAl + Fe3+ does not exceed 6.0 atoms per formula unit(and indeed rarely approaches this level).

Dioctahedral and trioctahedral 2:1 layered clays canbe considered to have the following general formulas:

Dioctahedral(EX^<MFe^UMg^‰Sig^A‰<OHXi (1)

e.g., montmorillonite: (1/2Ca,Na)066(Al3 34, M‰ 66)Si8O20(OH)4

Trioctahedral (EX+)e(Al,Fe3 + ̂ (Mg.Fe2 + )6-mSi8-sAl502o(OH)4 (2)

e.g., saponite: (1/2Ca,Na)066Mg6Si7.34Al066O20(OH)4

The exchangeable inter-layer cations are shown asmonovalent ions, EX+ . Structural formulas can be de-rived from electron-microprobe analyses in which wateris not determined directly, by assuming a total cationcharge of +44 in accordance with the anion comple-ment 02o(OH)4.

Dioctahedral and trioctahedral clays may be distin-guished by using their total of Si + Al + Mg + Fe*atoms, termed SAMF in the following discussion. Withreference to formulas (1) and (2), SAMF atoms for di-octahedral and trioctahedral clays are 12 and 14, respec-tively. The intermediate values between these extremesrepresent various mixtures of di- and trioctahedral claysin amounts proportional to SAMF. For example, a claywith a SAMF of 13 would have a composition corre-sponding to equal amounts of di- and trioctahedralclays, which could be either regularly or irregularlyinter-layered.

Alteration often proceeds in such a way as to favorthe formation of mixed-layer clays in preference to pureend-member di- or trioctahedral 2:1 layered clays. Con-sequently, in dealing with clay analyses, a general for-mula summarizing all possible 2:1 layer clays, irrespec-tive of SAMF, allows for a more realistic approach.Such a formula may be written as

with

)m(Mg)Fe2+)(4+2,.m)Si8.JAls02o(OH)4 (3)

SAMF = m + 8 + (4 + It - m) = 12 + It

where t is a measure of di-, trioctahedral, or intermed-iate character, such that t = O for a dioctahedral 2:1layered clay and t - 1 for a trioctahedral 2:1 layeredclay. By determining the charge balance from formula(3),

where

e + 3/n + 2(4 + It - m) + 4(8 - s) - 3s = 44

= 4-4t-m+s (4)

expression (4), relating the degree of elemental substitu-tion in octahedral and tetrahedral sites, and the inter-layer charge may be deduced for all 2:1 layered clays.

Three variables have been chosen from this relation-ship to give the maximum information about the struc-ture of 2:1 layered clays when plotted as a three-dimen-

552

CHEMISTRY OF ALTERATION MINERALS

sional graph. These three variables are inter-layer charge(e), total Al + Fe 3 + (m + s), and SAMF as a function oft in formula (4).

When t = O in formula (4) (dioctahedral 2:1 layeredclay), a series of values for (EX+) and total Al + Fe3 +can be calculated which satisfy the formula and take ac-count of the other constraints mentioned above, that(EX+) < 2 and total Al + Fe3 + < 6. Graphically, thesevalues delineate an inverted triangle (Fig. 1) enclosing afield in which dioctahedral 2:1 layered clays must fall.An opposing triangle with a common side defines thefield of trioctahedral 2:1 layered clay when t = 1 in for-mula (4). Several common alteration minerals have beenplotted on Figure 1 to aid in orienting the graph, andfields have been outlined to show the variation and dis-

tribution of members of the three groups of 2:1 layeredclays.

Fields can be constructed graphically for any value ofSAMF between the extremes of 12 and 14 using formula(4), each defining an area in which analyses satisfy astructural formula corresponding to a specific mixtureof di- and trioctahedral clays. These fields can be dis-played simultaneously using SAMF as a third axis in anisometric projection (Fig. 2).

The presentation of large numbers of clay analysesusing a method which simultaneously aids in their rapididentification has been a growing problem since the elec-tron-microprobe first became widely available. The meth-od outlined above is a possible solution to this problem.Figure 3 is a graph of (EX + ) versus total Al + Fe3 + of

Trioctahedral clay field

Dioctahedral clay field

phlogopite

Total Al + Fβ°

Figure 1. Fields of trioctahedral (SAMF = 14) and dioctahedral (SAMF = 12) 2:1 layered clays,calculated from formula (4), showing the distribution of common alteration minerals and clay groups.

Figure 2. Three-dimensional isometric projection showing the changes in field shape and size for varying mixturesof the end-member trioctahedral and dioctahedral 2:1 layered clays.

553

A. C. ADAMSON

6.0

5.0

4.0

3.0

2.0

1.0

1.0 2.0

Total Al + Fe

3.0 4.03+

Figure 3. Characteristics of clays from Holes 504A, 504B, 505, and5O5B. Relationships are obscured by the volume of data.

over 300 microprobe analyses in which the sheer volumeof data obscures any relationships. Even if these wereplotted as various weight percent oxides, relationshipswould be similarly lost. By adding a third dimension asa measure of trioctahedral character to the graph in Fig-ure 3, this problem can be overcome by effectively dilut-ing the results. In addition, analyses which comprise ofa variety of di- and trioctahedral clays and differing mix-tures, previously indistinguishable using a two-dimen-sional graph, become spatially oriented and allow real-istic comparisons between similar analyses. Clays whichfail to satisfy the field constraints calculated for theirparticular value of SAMF are also evident, drawing anatural division between analyses which are 2:1 layeredclays and those of some different type.

In practice, analyses are processed by computer. Ox-ide percents are recalculated on the basis of 22 oxygensand (EX+), total Al + Fe3+ , and SAMF determined foreach analysis. A program then divides the data accord-ing to SAMF, presently into eight groups between theextremes of 11.9 and 14.2, before final plotting in aseries of graphs of (EX+) versus total Al + Fe3+. Oneach graph is a set of clusters, each composed of similaranalyses (Table 1). Particularly in the case of the trioc-tahedral 2:1 layered clays (SAMF = 14), division intoclusters is made on the basis of structural formulas,because of the concentration of points. Dividing linescorrespond to the upper and lower limits of acceptable

substitution in known 2:1 layered-clay groups. Such di-visions are reinforced when other graphs with an in-creased dioctahedral component are considered, clustershaving begun to move apart (Fig. 4). For ease of pres-

4.0

×ui

Oδ 2.0

1.0

0.0

SAMF 13.00-13.50E-type

T-type

S-type

A-type

] Q-type

ö R-type

0.0 1.0

Tot

SAMF 13.51-13.80

2.0 3.03+

4.0

Ui

δ!•i.o

0.0 I0.0 1.0 2.0 3.0

3+4.0

Total Al + Fe

SAMF 13.81-14.10

Total Al + Fe'

Figure 4. The same analyses as in Figure 3, divided on the basis oftheir SAMF values. Relationships become apparent, similar analy-ses forming clusters. Shaded fields are drawn on the basis of struc-tural considerations, and are to some degree flexible.

554

CHEMISTRY OF ALTERATION MINERALS

entation, mean compositions are calculated for eachcluster (Table 2), and then plotted in a three-dimen-sional isometric projection (Fig. 5). The result is a seriesof trends, their position within the projection a functionof composition and clay type (Fig. 1). Direction indi-cates systematic changes in inter-layer charge and octa-hedral and tetrahedral substitution with changing SAMF.

Unfortunately, a three-dimensional projection suchas this is inherently difficult to use for determining thecoordinate values for a point in space. To overcome thisproblem, a standard graph of (EX+) versus total Al +Fe3+ (Fig. 6) is used in conjunction with a three-dimen-sional projection, both allowing easy reading of valuesand providing information on spatial relationships.

OXIDATION STATE OF IRON

The oxidation state of the iron has a marked effect onthe calculation of total Al + Fe3 + , and a less-marked ef-fect on other parameters. It cannot be determined by theelectron microprobe, and in calculating structural for-mulas it must be assumed. Wet chemical analyses ofsome samples of clay from Leg 69 give Fe2+:Fe3+ ratiosof about 1, though there are samples from oxidizedzones in which Fe3+ is dominant (Pertsev and Boroni-khin, this volume). Nevertheless, the calculations herehave been made with Fe entirely taken as ferrous ironfor the sake of simplicity and comparability.

Though the effect of this is small when iron concen-trations are low, it becomes important for iron-richclays. For typical samples of clays analyzed here, withtotal Fe about 1.2 atoms per formula unit, oxidation ofall of the iron would decrease EX+, SAMF, and total Alby about 3% relative, and would decrease (Mg + Fe2+)and increase (total Al + Fe3+) by the amount of ironinvolved (Table 3). The result on Figure 5 is to displacepoints to the right and slightly downward. The effect onmany analyses is small, but it may bring some analysesplotting to the left of the field of 2:1 clay minerals sothat they plot within it. This effect should be remem-bered when using Figure 5.

SECONDARY MINERALS

Clays

In hand specimen, the clays are of various shades ofgreen, lining or completely filling vesicles and veins, re-placing olivine and glass in both chilled pillow marginsand interstitial areas, and acting as cement to basalticbreccias. In zones of intense oxidation, red, orange, oryellow clays often envelope, partially or completely,green smectite cores. Crystallization of carbonates orvarious zeolites was a final phase of secondary-mineralformation: these phases form the centers of veins andvesicles, and more rarely pseudomorphs. Clays are ofseveral different morphological types. Fibrous varietiesare composed of a mass of elongate crystals which gen-erally grow perpendicular to fracture walls or aroundnuclei to form radiating fans and spheres. Layeredforms comprise of a series of different-colored claybands closely following the contours of the basalt walls.

Granular varieties appear mottled, composed of clays ofa single color but of differing shades. Amorphous formsshow no structure; they are usually dark in hand speci-men and opaque when viewed in thin-section.

In thin-section, clays are usually various shades ofbrown and green, but may be yellow, orange, and red inoxidized zones. Clay morphologies in thin-section varyin the same way as in hand specimen. Several differentforms in a single vein can be used as indicators of theconditions prevailing at time of precipitation. It wouldseem reasonable to suppose that good crystal shape de-velops under stable and prolonged conditions of freeuninhibited growth, and irregular crystal forms underconditions favoring more rapid formation. Layering,particularly in veins, must be a consequence of fluidfluctuations during clay precipitation, and, though op-tically apparent, internal geochemical variations areoften absent. Combinations of the various morpholo-gies often allow for the determination of a complexcrystallization history, which is then capable of being in-vestigated geochemically.

In particularly large cavities and veins, clays formisolated, radiating spheres, occasionally showing layer-ing. These eventually merge when in sufficient abun-dance to form an amorphous or plumose mass. This ra-diating morphology would suggest free growth around anucleus, probably in the form of small fragments fromclay-coated walls.

Clays forming olivine pseudomorphs follow a char-acteristic veining pattern as a result of the initial altera-tion down cracks. Depending on local conditions, claysare either highly birefringent fibrous varieties, occasion-ally containing blebs rich in sodium and calcium, ordark, isotropic forms.

The alteration of glass varies in intensity between dif-ferent areas of the core, but in general pale-brown orgreen clays replace pillow margins with veining by zeo-lites, and darker clays interstitial glass, unless stainedbright orange in oxidized areas. The alteration of glassis an in situ process, and the minerals which are formedare fundamentally different from those produced byprecipitation in veins and vesicles. Geochemically, themajority are from a single montmorillonite-like min-erals group rich in the less-mobile elements such as Aland Ti. In contrast, the clays precipitated in veins andvesicles are more varied in their composition and claytype.

Geochemistry

Geochemically, these analyses form six distinct trendswhen plotted on a three-dimensional graph using themethod and techniques described earlier (Figs. 5 and 6).

The geochemistry and petrological significance ofthese six groups will be considered in two sections, thefirst covering the 2:1 clay types T (talc), S (saponite),and A (montmorillonite-like minerals), and the second,clay mixtures containing an end-member component withan unknown compositional formula (here labelled R, Q,and E minerals, as described below).

555

A. C. ADAMSON

Table 1. Selected microprobe analyses of clay minerals from Holes 504A, 504B, 505, and 5O5B.

Analysis No.

Sample(interval in cm)

Typea

310

505B-2-1,25-28

E2

173

505B-2-2,100-103

E2

453

504B-40-2,89-91

E3

307

505B-2-1,25-28

E3

349

5O5B-3-1,109-112

TO

356

5O5B-3-1,109-112

TO

289

5O4B-61-3,1-4

Ti

440

504B-63-4,65-69

Ti

403

5O5B-5-1,50-60

Ti

239

5O4B-58-1,35-38

so

135

504B-21-4,120-124

so

30

504B-7-3,27-30

Si

SiO2

TiO2

A12O3

FeOMnOMgOCaO

K2OTotal

SiAlAl

Fe2 +

MnMgCaNaK

Mean SAMFTotal Al + Fe3 +

(EX + )(Mg + Fe2 + )

56.90

0.666.40

26.410.103.950.30

94.73

7.640.10

0.72

5.280.011.030.05

13.770.101.116.00

46.80

26.20

9.190.367.642.62

92.83

7.43

3.47

2.170.612.350.53

13.07

3.005.64

49.30

4.068.10

23.211.20

13.140.27

99.30

6.720.65

0.92

4.710.173.470.05

13.010.653.875.64

54.20

6.00

26.100.15

10.620.28

97.30

7.32

0.68

5.260.022.780.05

13.25

2.875.93

46.30

1.2320.10

17.190.391.991.34

88.57

7.260.23

2.63

4.010.070.600.27

14.130.231.006.64

43.60

1.3319.80

13.800.340.790.18

81.65

7.48

0.27

2.84

3.530.060.260.04

14.120.270.436.37

54.40

1.6714.20

20.171.13

0.18

91.74

7.720.28

1.68

4.270.17

0.03

13.950.280.375.95

55.50

1.2010.90

23.090.80

0.12

91.58

7.75

0.20

1.27

4.800.12

0.02

14.020.200.266.07

54.10

0.819.50

22.080.22

0.29

87.21

7.870.140.01

1.16

4.78

0.03

0.05

13.950.140.125.94

39.40

5.8118.20

13.952.130.57

80.29

6.751.17

2.61

3.560.390.19

14.101.170.976.17

45.80

5.068.90

20.960.700.940.25

82.61

7.140.93

1.16

4.870.190.280.05

14.110.930.576.04

47.80

5.7211.20

20.520.691.370.51

87.86

7.090.910.09

1.38

4.540.110.390.10

14.011.000.715.92

Letters refer to mineral type (see text); subscript numbers indicate the SAMF division to which the analysis is assigned, as follows: 6, SAMF < 12.00; 5,12.00-12.50; 4,12.51-13.00; 3,13.01-13.50; 2,13.51-13.80; 1,13.81-14.10; 0, SAMF < 14.10.

2:1 Clay Minerals

Talc-like Minerals

The graphical position of pure talc, with composi-tional formula Mg6Si8O20(OH)4, is shown in Figure 1.The 31 analyses representing the clay types Tj are chem-ically very close to this ideal composition, and can al-most be considered as pure talc. They have an averageSAMF of 14, and a formula only slightly different fromthat of the pure mineral,

This difference arises from the limited substitution ofsilica by aluminium in tetrahedral sites, and by a corre-sponding small increase in the inter-layer charge.

Although the chemistry of T! minerals is similar tothat of talc, no talc has been identified positively bymeans of XRD in these rocks. In addition, no opticalidentification has been confirmed during this study.Possibly these analyses are of an unusual mineral typewith a very poor exchangeable-cation capacity.

A line connecting To, Tlf and T2 (Fig. 5) indicates theapproximate plotting position of any analysis with atalc-like composition between possible extremes ofSAMF. A lower limit of SAMF = 13.6 is evident fromavailable analyses; below this, talc-like minerals wouldappear not to form. In the case of the To analyses, anaverage SAMF of 14.3 is above the maximum for atrioctahedral clay. These analyses are minerals poor inexchangeable cations, but rich in the cations of whichSAMF is a measure.

Particularly in the case of the Tj analyses, separationfrom the more-saponitic minerals, Sj, has been made on

the basis of structural formulas, allowing a certain flexi-bility in the position of the dividing line. Analyses closeto this line may well be mixtures of talc-like mineralsand saponites in varying proportions. Once SAMF hasdecreased to around 13.8, clusters have diverged suf-ficiently to eliminate any such trend (Fig. 4).

Talc-like minerals are most common as the alterationproduct of olivine, particularly in the lower section ofHole 504B. A few examples are present at Site 505, butgenerally the olivine is reasonably fresh here comparedto other sites. Most pseudomorphs are homogeneous incomposition. Only one heterogeneous olivine pseudo-morph has been found containing talc-like mineralsmixed with other phases (Sample 504B-41-2, Piece 605).In vesicles, minerals other than the talc-like phases areoften present, acting as either rims or cores.

Saponites

Clays with a compositional formula corresponding tothat of saponite, or some variant of its basic formula,are by far the most common alteration minerals presentin basalts from the Costa Rica Rift.

Trioctahedral saponites in these rocks, Sj, have aderived formula,

(EX+)0.77(Mg,Fe2 + )6.oSi7.1Al0.9020(OH)4.

They differ only slightly from an ideal saponite, wheresubstitution in both inter-layer and tetrahedral sites is0.66. In the other S-type minerals conforming to thisbasic formula, substitution by aluminium in octahedralsites becomes more pronounced as SAMF values de-crease, and the dioctahedral component of mixtures be-comes dominant. The increase in inter-layer charge from

556

CHEMISTRY OF ALTERATION MINERALS

Table 1. (Continued).

242

504B-58-1,35-38

Si

408

504B-36-2,121-123

Si

369

504B-54-1,31-34

S2

121

504A-6-3,12-15

S2

933

504B-14-2,1-17

S2

921

504B-9-2,11-14

S3

286

504B-49-1,51-54

S3

112

504A-6-3,12-15

S3

203

504B-7-2,98-102

Al

155

504B-37-1,15-17

Al

232

504B-69-1,106-112

Al

924

504B-6-1,33-43

R3

222

504B-28-5,19-22

R3

41.50

6.3317.60

15.232.810.760.10

84.65

6.711.21

2.38

3.670.490.240.02

13.971.211.246.05

48.800.133.38

14.90

18.381.630.930.22

88.38

7.330.60

0.011.88

4.120.260.270.04

13.920.600.845.99

48.70

7.6814.40

18.873.141.700.18

94.72

6.851.27

1.70

3.960.470.460.03

13.791.271.445.66

52.10

4.4914.80

16.081.231.141.85

91.75

7.540.460.30

1.79

3.460.190.320.34

13.560.761.045.56

50.800.277.08

10.900.08

20.331.811.730.09

93.10

7.090.910.970.031.270.014.230.270.470.17

13.751.161.035.50

49.500.398.66

10.900.08

16.351.801.941.02

90.65

7.120.880.590.041.310.013.500.280.540.19

13.411.471.284.82

48.00

5.9516.90

14.903.350.722.26

92.35

7.090.910.13

2.09

3.280.530.210.43

13.491.031.695.37

52.90

3.7717.80

13.421.070.961.53

91.46

7.740.260.39

2.18

2.960.180.270.29

13.490.650.895.10

44.900.25

13.3216.900.19

12.560.85

1.97

90.91

6.630.142.18

2.090.032.770.14

0.37

13.802.320.644.85

44.50

7.787.60

21.302.670.890.26

84.98

6.761.240.15

0.97

4.830.440.260.05

13.951.391.195.79

43.90

8.5316.50

15.602.600.470.13

87.80

6.741.260.28

2.11

3.580.430.140.03

13.981.541.025.70

48.800.358.76

10.600.12

17.584.491.880.16

92.97

6.901.100.360.041.260.013.700.680.520.23

13.321.461.904.96

43.200.889.03

23.80

9.412.673.390.27

92.66

6.611.390.240.103.04

2.140.441.000.05

13.411.631.935.18

Sj to S4 (Table 1) is principally due to an increase in K+

content.Unlike the talc-like minerals, different saponites com-

monly occur together, and they exhibit a wide range ofmorphologies irrespective of their location. They oc-cupy the complete spectrum of alteration sites: veins,vesicles, pseudomorphs, matrix, and marginal glassareas. In many cases, saponites filling vesicles and veinsare stratified with interleaved layers of other mineraltypes. There appears to be no relationship linking theorder in which layers of different chemistry occur, ex-cept that SAMF values are often lower in border areasthan in the central parts of veins and vesicles. In severalcases, saponite-filled vesicles have a core composed ofSt clays and a rim of S3, a transition requiring a decreasein SAMF but an increase of K+ in inter-layer sites.

Montmorillonite-like MineralsThe clay analyses belonging to this series can be di-

vided into two groups, one resembling a very aluminoussaponite (Aj-A^, and the other more akin to an idealmontmorillonite mineral (A<̂ -Ag) (Table 2).

In the first group (Aj-A^, the increases in aluminiumare accommodated in octahedral sites. It would seemhighly probable that the analyses belonging to types Ajand A2, are mixtures of saponites and the montmoril-lonite-like minerals, A^Ag, in varying proportions. Dis-tinct clusters are evident of particular mixtures, how-ever, and I feel that these warrant treatment as separategroups. In the second A-type group (A4-A6), analysesresemble ideal montmorillonite and plot well away fromsaponitic minerals, mixing of the two having finishedwith the A2 minerals. The closest any of these analysescomes to an ideal montmorillonite is:

(EX+)o.9(Al,Fe3 + ,Ti)2.7(Mg,Fe2+ .

There are significant amounts of titanium and manga-nese in these analyses.

The purer dioctahedral clay members of this series,notably A^Ag, rich in the less-mobile elements such asAl and Ti, are formed by the in situ alteration of glass,principally in pillow margins. There are no occurrencesof these minerals constituting pseudomorphs, and onlyrare instances as fillings of vesicles. Altered glass mar-gins are often composed of a series of fragments formedduring cooling, separated by layered veins. The frag-ments are optically layered, but chemically are relativelyhomogeneous, composed entirely of aluminous clays oftypes A4 and A5. The veins are more diverse, composedof a variety of clay types. The most common are theminerals Aj-A3, but there are occasional layers of sap-onite, S1} and rare phases such as R3. The association ofsaponite with A!~A3 minerals illustrates the mixing be-tween these groups, discussed earlier. Saponites precipi-tated in the marginal areas of veins adjacent to mont-morillonite-like clays of types A^Ag would appear tohave an effect of locally diluting these to form mineralsof types A!-A3. Formation of zeolites in the centralparts of veins is the final phase of precipitation.

Microprobe analyses of altered interstitial matrix glassaway from areas of intense replacement shows alter-ation processes to be similar to those active within pil-low margins, favoring formation of aluminum-rich clays.These are mainly A and A2-type clays, but a few Sisaponites and rare R- and E-type minerals occur.

Non-2:1 Clay Minerals

The three secondary mineral series designated E, Q,and R are not 2:1 layered clays, but may represent eithermixtures of various 2:1 layered clays with an end mem-ber of uncertain composition, or alternatively a separateseries of secondary clay minerals, different from 2:1

557

A. C. ADAMSON

layered clays, but with their own characteristics andtrends.

Group R Minerals

This group contains analyses which appear to repre-sent varying mixtures of a 2:1 layered clay and an un-identified mineral of R5 type, rich in Na and Ca. Inter-layer charge is characteristically high, exceeding 2 in thecase of the mineral types R4 and R5. Analyses formingthe two types R2 and R3 have an inter-layer charge closeto 2, and although chemically they meet the require-ments for a 2:1 layered clay, derived compositional for-mulas resemble no common 2:1 layered clay mineral.

Analyses belonging to this series are not common andare restricted to vein and vesicle fillings and to unde-fined matrix material in intensely altered breccias. Thetwo mineral types R2 and R3, which resemble clays, arefound exclusively as layers in vesicles and very occasion-ally in veins. They can act as the outer rim of a layeredseries, saponite (S!-S3) and talc forming the core. Theanalyses which correspond to types R4 and R5 are frombreccias and amorphous isotropic areas of alteration inthe basalt matrix. It seems probable that in such casescontamination by varying amounts of whole rock mayhave occurred.

Group Q Minerals

The main characteristic of this group is the very highconcentration of inter-layer charge, which excludes themfrom being 2:1 layered clays. Calcium is an importantcomponent, varying from 7 to 14 oxide percent in typesQ3 and Q5, respectively, considerably more than in anyother alteration mineral found in these rocks. Theamounts of aluminum and silica fluctuate only slightlythroughout the series; the decrease in SAMF results froman overall decrease in magnesium and iron, but followsno simple pattern. Iron can increase at the expense ofmagnesium or vice versa, while still maintaining an over-all decrease in their totals from Q3 to Q5.

Minerals with this chemistry occur in veins, pseudo-morphs, and vesicles. In the case of the latter, no layer-ing is evident, the vesicle apparently being filled by asingle phase. Other vesicles in the same sample may con-tain different clay minerals; the saponite type S^ for ex-ample, can be found in close proximity to minerals oftype Q.

When these minerals are one of several phases filling apseudomorph, some interesting relationships occur.Generally they form isolated blebs in the center of highlybirefringent clays of different chemistry, principallysaponites and talcs. Blebs are usually composed of oneof the two mineral types Q4 and Q5, but in one case anenvelope of E5 has formed. The latter is a highly sodicmineral. If the blebs are a later phase of alteration, thenthe process which differentially concentrates calcium in-to these small areas, and iron in the case of Q4, ap-parently does not affect surrounding talcs or saponites.Possibly they are the products of a process of dif-ferentiation or segregation operative during simultane-ous formation of the two phases, magnesium preferen-

tially being absorbed into saponite or talc minerals, andcalcium, and to a lesser extent aluminium, into theblebs.

Two examples of olivine pseudomorphs completelyfilled by secondary minerals of composition Q4 havebeen recorded. They are opaque in thin-section, prob-ably a result of their high iron content.

Group £ Minerals

The most notable characteristic of the analyses form-ing this group, is their diversity in sodium content fromone end of the series to the other. Expressed as oxidepercents, sodium forms only 0.6 of the Eo mineral analy-ses, but increases to 14.0 in the case of mineral E5.

If iron in these analyses is assumed to be completelyreduced, the Eo-E3 minerals with an inter-layer chargeless than 2 (Figs. 5 and 6) fail to satisfy the field require-ments for 2:1 layered clays, and therefore must be struc-turally of some different mineral type. However, oncesome degree of partial oxidation is assumed in the Eo-E3minerals, spatial movement is sufficient to bring theseanalyses within the 2:1 layered clay fields. Althoughstructurally these analyses then conform to the require-ments for a 2:1 layered clay, their derived formulas donot resemble any common clay mineral.

Analyses of the two most sodic members of the seriesare presently limited to one of each. Exchangeable ca-tions far exceed the acceptable maximum for clays, anda ratio of (Na + K):(Mg + Fe):Si is probably the mostuseful means of illustrating relative element abundancesfor comparison with the other mineral groups. In thecase of the mineral E4, this ratio is 4:6:7, and for E5 it is6:6:6.5. Whether these two analyses represent an end-member mineral which is mixed in varying proportionswith 2:1 layered clays to form the E-type trend is unclearat this time.

Members of the group with low inter-layer charge,Ej-E3, when occurring together, form a continuous ser-ies, particlarly when filling layered vesicles. In suchcases, there is an increase in sodium from the rim, El5inward through an inner ring of E2, and finally to a corecomposed of the sodium-rich mineral E3.

Pseudomorphs are the source of most analyses of thistype. They generally occur with other secondary phasesas rims or internal veins, but may very rarely form thehigh-sodium, high-calcium blebs mentioned in an earliersection.

Chlorite Mixed-Layer Minerals

The presence of chlorite-smectite mixed-layer min-erals is based primarily on geochemical evidence, withsome additional support from diffractogram sources.

An ideal chlorite analysis calculated on the basis of22 oxygens has a zero exchangeable cation capacity, anda SAMF value of 15.7. Dilution of chlorite by varyingamounts of 2:1 layered clays increases the inter-layercharge from zero to values proportional to the percent-age of clay in the mixture. SAMF decreases from 15.7 ina similar manner. The principal criteria used in recog-

558

Table 2. Average chemical compositions for all clusters.

Type3

Siθ2TiO2

AI2C3FeOMnOMgOCaONa2θK2O

Total

SiAlAlT i ,F e 2 +MnMgCaNaK

No. of AnalysesMean SAMFTotal Al + Fe3 +(EX + )(Mg + Fe 2 + )

E 0

46.20

0.1222.60

15.030.330.611.61

86.50

7.480.02

3.06

3.630.060.190.33

414.200.020.646.69

El

52.80

1.5911.70

22.730.612.030.35

91.81

7.500.27

1.39

4.810.090.560.06

1214.000.270.806.20

E 2

54.00

1.087.380.02

25.340.245.140.29

93.49

7.460.17

0.850.015.220.031.380.05

913.700.171.496.10

E3

52.10

0.747.46

24.400.27

10.670.30

95.94

7.210.12

0.86

5.030.042.860.05

913.220.122.995.89

E 4

43.10

1.4728.10

5.010.289.113.84

90.91

7.220.29

3.94

1.250.052.960.82

112.700.293.785.20

E 5

38.20

3.8216.30

10.322.34

14.120.31

85.41

6.540.77

2.33

2.630.434.680.07

112.270.775.614.96

TO

44.80

1.1823.30

0.0315.200.530.521.04

86.60

7.270.23

3.160.013.680.090.160.21

514.330.230.556.84

Ti

53.800.021.52

10.85

22.080.730.300.16

89.45

7.710.3

0.011.30

4.720.110.080.03

3113.990.30.336.0

T 2

52.600.050.345.360.07

22.100.310.400.34

81.57

8.01

0.060.010.680.015.020.050.120.07

1213.770.060.285.7

so44.10

0.015.88

13.820.02

19.870.941.130.19

85.96

6.831.07

1.790.014.590.160.340.04

1414.28

1.070.696.4

Si

47.000.075.36

11.54

20.071.191.160.25

86.64

7.100.900.05

1.46

4.520.190.340.05

6114.030.950.775.98

S2

47.700.095.87

13.780.02

16.742.251.220.41

88.07

7.170.830.210.011.73

3.750.360.360.08

2413.69

1.041.165.48

S3

50.700.125.10

17.730.04

12.201.720.952.53

91.09

7.530.470.420.012.200.012.700.270.270.48

2113.320.891.294.90

S 4

52.30

3.0324.43

5.530.920.335.91

92.45

8.00

0.55

3.13

1.260.150.101.15

412.940.551.554.4

Al

46.100.23

10.1411.610.10

18.551.730.790.59

89.84

6.721 ">80.480.021.410.014.030.270.220.11

1213.90

1.740.875.44

A2

49.300.329.98

11.620.66

17.571.841.260.25

92.80

6.931.070.610.031.370.083.680.280.340.05

713.63

1.650.955.05

A3

44.000.91

15.0214.400.24

11.441.201.481.49

90.18

6.501.501.220.101.780.032.520.190.420.28

513.422.621.084.30

A4

52.000.39

23.286.400.065.271.480.881.26

91.02

7.050.952.810.040.730.011.060.220.230.22

212.54

3.720.881.79

A 5

54.100.40

23.146.200.094.781.441.061.42

92.63

7.080.922.680.040.540.010.930.200.270.24

412.103.560.911.47

A 6

58.20

23.731.60

0.191.603.880.13

89.33

7.730.273.45

0.18

0.040.231.000.02

311.66

3.721.480.21

R2

45.30

8.6210.40

19.335.700.52-0.33

90.20

6.641.360.13

1.27

4.220.890.150.06

213.62

1.491.995.49

R3

49.600.279.26

11.820.07

17.313.722.020.43

94.50

6.911.090.430.031.380.013.590.550.540.08

913.40

1.521.724.97

R4

48.900.38

12.3310.470.08

11.604.153.200.38

91.49

6.981.021.060.041.250.012.470.630.880.07

512.782.082.213.72

R5

53.900.67

11.7410.800.04

10.035.673.250.30

96.40

7.270.731.210.071.220.012.020.820.850.05

312.38

1.872.543.24

Q3

43.100.168.239.50

17.357.091.060.20

86.69

6.601 400.090.021.22

3.961.160.310.04

613.27

1.492.675.18

Q4

36.500.747.78

23.300.088.17

11.360.300.36

88.59

6.071.52

0.103.240.012.022.020.100.07

512.85

1.524.215.26

Q5

46.700.77

10.3712.500.15

10.6113.680.76

95.54

6.641 360.380.081.490.022.252.080.21

212.12

1.744.373.74

a See footnote under Table 1.

A. C. ADAMSON

SAM F< 12.00SAMF 12.00-12.50SAMF 12.51-13.00

13.01-13.5013.51-13.8013.81-14.10

SAMFSAMFSAMFSAMF>14.10

© E-typeΦ T-typeQ S-typeΔ A-type

Q-typeR-type

Figure 5. Average values calculated for each cluster of analyses in Figure 4 and plotted on a three-dimensionalisometric projection. Fields for each division of SAMF are shown, as in Figure 2. Points belonging to asingle mineral type are joined by continuous lines to show the overall spatial relationships between the dif-ferent mineral types. Note particularly how most of the E-, Q-, and R-type minerals fail to satisfy the fieldrequirements for the 2:1 layered clays.

nizing chlorite analyses are high SAMF values with lowinter-layer charge.

The most convincing mixed chlorite analysis availa-ble at present comes from the center of a layered veincomposed of A-type minerals, in the altered glass mar-gins of a pillow (Sample 504B-7-2, Piece 466). The chem-istry can be explained by mixing in equal amounts sapo-nite with an approximate formula (EX+)04 (Mg,Fe2+)6

Si7 6Al0 402o(OH)4 and a chlorite such as pynochlorite(Deer et al., 1962) Al25(Mg,Fe2+)9.5Si5.7Al2.7O20(OH)16

It is of interest to note that the chlorite, characteristicof higher temperatures of alteration, is found close tothe basalt/sediment interface, where temperatures arerelatively low. Precipitation of this phase formed thefinal episode of alteration in the pillow basalts at thishorizon.

Zeolites

At the time of this writing, little work has been con-ducted on zeolites. Thomsonite and analcite have beenpositively identified from Hole 504B by XRD methods.These, in common with other zeolite occurrences, fromvisual and microscopic observations of the core, occureither alone or with clays in varying proportions as fill-ings to veins and pseudomorphs. Zeolites also appear inmany pillow margins, forming relatively late in the al-teration sequence. The growth of fracture-filling zeo-lites usually follows an earlier phase of clay precipita-tion, giving a layered appearance, characteristic radiat-

ing crystal sprays completely filling veins where perme-ability fell sufficiently to inhibit smectite formation(Velde, 1977). In other cases, the contemporaneous de-velopment of zeolites and other minerals diminishes lay-ered appearances and results in a mottled effect.

Carbonates

Both aragonite and calcite have been positively iden-tified in several samples from Hole 504B using XRD. Ofthese samples, the deepest is approximately 150 metersinto basement, though carbonates are usually evident tomuch greater depths than this.

Carbonates appear late in the alteration sequence,forming the centers of veins, green smectites generallyacting as an envelope against the rock walls. Contactsare sharp, lacking any gradational zone. Aragonite crys-tals often cut or truncate smectite fibers growing per-pendicular to vein walls. In other cases, crystallizationof aragonite follows existing smectite layers, filling theremaining cavities. Occasionally, a vague discontinuityis evident down the center of calcite veins filled by amor-phous clay material, probably the result of renewed frac-turing and late smectite precipitation.

Following the reasoning of Pritchard (1979), rapid al-teration of basalt glass at a late stage seems to providenot only calcium ions, but also hydroxyl ions. These hy-droxyl ions displace the normal bicarbonate-carbonateequilibrium in such a way that large quantities of calcitecan form. This process is probably assisted by the re-

560

12.00

12.00-12.50

12.51-13.00

13.01-13.50

13.51-13.80

13.81-14.10

14.10

2.0

Total Al + Fc

Figure 6. The isometric projection of Figure 5 projected onto a singleplane, the SAMF divisions to which each point belongs, and thenumber of contributing analyses in the original cluster (Fig. 4).(Symbols same as in Fig. 5; numbers in italic denote number ofanalyses for that point.)

CHEMISTRY OF ALTERATION MINERALS

moval of Mg ions from solution, into clays, which nor-mally inhibit calcite nucleation.

DISCUSSIONThe alteration processes replacing primary minerals

in situ and those precipitating clays in veins and vesiclesdiffer substantially. This section will first discuss theprocesses which control the precipitation of clay withinvesicles, and then those which control the filling of frac-tures. Finally, the processes affecting in situ alterationwill be briefly discussed.

Individual mineral trends within vesicles recognizedby using three-dimensional graphics probably delineatesome form of continuous reaction series. Possibly ionicdiffusion, or changes in circulating-fluid chemistry canaffect certain tri- and dioctahedral minerals in such away as to systematically change their chemistry to fol-low a specific mineral trend. This continuous reactionseries is especially evident in the case of the saponitesand the E-type minerals. In both cases, optically layeredvesicles occur which contain several members of a par-ticular series in sequence, each with diffuse contacts. Toaccount for these observations, two possible processesof formation may be operating—first is a process of"constructive" alteration; here, very slow but continu-ous precipitation of a mineral phase occurs, graduallychanging chemically to reflect the fluctuations in circu-lating fluid chemistry with time.

Second is a process of "destructive" alteration whereleaching processes are dominant. Pre-existing mineralsare chemically changed to follow one of the chemicaltrends by removal of mobile elements into circulatingpore fluids. In the case of layered vesicles containingminerals from different series with sharp contacts, itseems likely that a different process of formation maybe operative. In contrast to the envisaged continuous

Table 3. Effect of oxidation state of iron on graphical parameter.

Sample (interval in cm)

Typea

Analysis

SiO2

TiO2A12O3

Fe2O3

FeOMnOMgOCaONa2OK2O

Total

SiAlAlT i ,Fe3 +

Fe2 +

MgMnCaNaK

SAMFTotal Al + Fe3 +Total Al(EX + )(Mg + Fe2 + )

504B-6-1,

Microprobe

48.800.358.76

10.100.12

17.584.491.880.16

92.97

6.901.100.360.04

1.263.700.010.680.510.03

13.321.461.461.904.96

33-34R3

Wet Chemical

5.204.90

6.861.140.310.030.550.573.680.020.670.510.03

13.112.001.451.884.25

504B-29-1,

Microprobe"

49.10

11.24

11.00

17.381.631.200.15

91.63

6.921.080.79

1.293.65

0.250.330.03

13.701.871.870.854.94

11-14A 2

Microprobe0

11.00

6.781.220.60

1.14

3.58

0.240.310.02

13.322.961.820.813.58

504B-7-3,

Microprobe*3

49.20

V93

10.0

21.050.580.570.74

86.13

7.380.620.07

1.264.70

0.090.160.14

14.030.690.690.495.96

27-30

SiMicroprobe

10.0

7.240.67

1.11

4.61

0.090.160.14

13.631.780.670.48

13.63

* See footnote under Table 1.b Total iron is assumed to be Fe2 + or Fe3 + .

561

A. C. ADAMSON

processes of formation which produce diffuse layering,sharp contacts suggest that cavities are filled by a multi-stage process. Initially a rim is precipitated, followed byformation of a core of a different mineral type.

Irrespective of the type of process precipitating lay-ered clays, their study can act as an indicator of tem-poral change in fluid chemistry in the area of vesicles.

Clay precipitation in veins appears to be subject toconsiderable variation. Clays of many different morph-ologies often occur in close proximity, and layering,when evident, can be sharp or diffuse. Late-phase car-bonates and zeolites are common in several localities.Initial investigation of layered veins has shown remark-ably little chemical variation between individual layersand clays of different morphology. Apparently, the pres-ence of visible layering bears little relationship to major-element chemistry. This will be further investigated. Mostveins away from pillow margins contain minerals belong-ing to a single clay type. Where veins are more hetero-geneous, precipitation is probably governed by the sameconsiderations discussed earlier for the case of vesicles.Possibly the lack of clay variety in most analyzed veins isa result of precipitation from pore fluids which are chem-ically highly constant. Diffusive alterations of pre-ex-isting secondary phases, termed destructive alterationabove, are probably of little importance in the filling ofveins.

The process of in situ alteration of primary mineralsand basaltic glass differs fundamentally from that gov-erning the precipitations of clay from solution. Here, pri-mary mineralogy, not pore-fluid chemistry, exerts a verystrong, if not dominant effect on the occurrence and dis-tribution of the secondary minerals. In the case of basal-tic glass, mineral phases develop which are rich in theless-mobile elements such as Al and Ti, forming themontmorillonite-like clays, A4-A6. Such phases are rare-ly found as the products of precipitation. Alteration ofolivine generally forms Mg-rich phases, principally talc-like minerals and saponite. The alteration of olivine be-gins at the rim and along cracks penetrating the crystal,and later proceeds to the rest of the crystal. For plagio-clases, the calcic core is the first part of a crystal to be re-placed, whereas the sodic rims consistently remain intact.

No consistent differences in the chemistry of the al-teration minerals can be found between the cold, opencrust at Site 505 and the hot, sealed crust at Sites 501and 504. The various secondary mineral types describedhere are present in the rocks from each hole. There is,however, a marked difference in the degree of replace-ment that has occurred. The hot, sealed crust is muchmore intensely altered than the cold, open site, wherefresh olivine can still be found. This difference can beassigned to the difference in temperature and the conse-quently different rates of alteration in the two areas.However, the difference in temperature does not appearto have been sufficient to affect the chemistry and min-eral type of the minerals being formed.

In many of the thin-sections probed in this study,particularly those from Site 504, there are a great num-ber of different clay types in close proximity. These maybe in homogeneous or heterogeneous vesicles, veins, or

pseudomorphs. Such a variety on a small scale suggeststhat the circulation of pore fluids within the rock is veryrestricted in scope. Rapid flow, producing a high water/rock ratio, would lead to the imposition of a homogen-eous alteration mineralogy on the whole rock, becausethe controlling chemistry would be that of the circulat-ing solution, not that of the local solids. Here it seemsthat water/rock ratios generally must have been small,so that the rock has dominated the alteration pattern,even in the hot, sealed crust at Sites 501 and 504.

At Site 504, the general pattern of a high variety ofthe clays developed persists with depth in the crust.Clearly, the higher temperatures at the bottom of thehole were not able to promote the circulation of water toa degree sufficient to affect the essentially rock-domi-nated chemistry. If free circulation is confined withinthe section to fault zones or other regions of intensecracking, sections of the core should be identifiable inwhich a water-dominated homogeneous clay associationdevelops. This study has not identified such an area, butthe samples were taken at individual spots down thecore not necessarily associated with fracturing. It is pos-sible that further study will identify water-dominatedregions in the core, and this work is in progress.

ACKNOWLEDGMENTS

Grateful thanks go to Pete Treloar of Cambridge for help in run-ning the electron-probe, to Peter Oakley for advising with wet chemi-cal analysis, to Dave Robson for his assistance in the computer handl-ing of data, and to Joe Cann for the original idea of presenting clayanalyses using three-dimensional graphics, and for many stimulatingdiscussions. I would also like to thank Dr. M. H. Battey and Dr. GaryPritchard for reviewing this paper. The author was supported by aN.E.R.C. grant throughout the work.

REFERENCES

Andrews, A. J., 1977. Low temperature fluid alteration of oceaniclayer 2 basalts, DSDP Leg 37. Can. J. Earth Sci., 14:911-926.

Bailey, S. W., 1980. Structures of layered silicates. In Brindley, G.W., and Brown, G. (Eds.), Crystal Structures of Clay Mineralsand Their X-Ray Identification: Mineralogical Society, Mono-graph No. 5, London, pp. 1-124

Baragar, W. R. A., Plant, A. G., Pringle, G. J., and Schaw, M.,1977. Petrology and alteration of selected units of Mid-AtlanticRidge basalts sampled from Sites 332 and 335, DSDP. Can. J.Earth Sci., 14:837-874.

Bass, M. N., 1976. Secondary minerals in oceanic basalts, with specialreference to Leg 34, Deep Sea Drilling Project. In Yeats, R. S.,Hart, S. R., et al., Init. Repts. DSDP, 34: Washington (U.S. Govt.Printing Office), 393-432.

Bohlke, J. K., Honnorez, J., and Honorez-Guerstein, B.-M., 1980.Alteration of basalts from Site 396B, DSDP: petrographic andmineralogic studies. Contr. Mineral. Petrol., 73:341-364.

Cann, J. R., 1979. Metamorphism in the oceanic crust, 1979. In Tal-wani, M., Harrison, C. G., and Hayes, D. E., (Eds.), Deep Drill-ing Results in the Atlantic Ocean: Ocean Crust: Washington (Am.Geophys. Union), pp. 230-238.

Deer, W. A., Howie, R. A., and Zussman, J., 1962. Rock FormingMinerals (Vol. 3): London (Longmans).

Donnelly, T. W., 1978. Low temperature alteration on the oceaniccrust [paper presented at the Ewing Symposium, New York,March, 1978].

Donnelly, T. W., Pritchard, R. G., Emmerman, R., and Puchelt, H.,1979. The aging of oceanic crust: synthesis of the mineralogicaland chemical results of Deep Sea Drilling Project, Legs 51 through53. In Donnelly, T. W., Francheteau, J., Bryan, W. B., Robin-son, P. T., Flower, M. F. I., Salisbury, M., et al., Init. Repts.DSDP, 51, 52, 53, Pt. 2: Washington (U.S. Govt. Printing Office),1563-1577.

562

CHEMISTRY OF ALTERATION MINERALS

Donnelly, T. W., Thompson, G., and Salisbury, M. H., 1979. Thechemistry of altered basalts at Site 417, Deep Sea Drilling Project.In Donnelly, T. W., Francheteau, J., Bryan, W. B., Robinson,P. T., Flower, M. F. I., Salisbury, M., et al., Init. Repts. DSDP,51, 52, 53, Pt. 2: Washington (U.S. Govt. Printing Office),1319-1330.

Hart, R. A., 1973. A model for chemical exchange in the basalt-sea-water system of oceanic layer 2. Can. J. Earth Sci., 10:799-816.

Pertsev, N. N., and Rusinov, V. L., 1979. Mineral assemblages andprocesses of alteration in basalts at Deep Sea Drilling Project Sites417 and 418. In Donnelly, T. W., Francheteau, J., Bryan, W. B.,Robinson, P. T., Flower, M. F. I., Salisbury, M., et al., Init.Repts. DSDP, 51, 52, 53, Pt. 2: Washington (U.S. Govt. PrintingOffice), 1219-1242.

Peterson, N., 1978. Low temperature alteration of the magnetic min-erals in ocean floor basalts [a paper presented at the Ewing Sym-posium, New York, March, 1978].

Pritchard, R. G., 1979. Alteration of basalts from Deep Sea DrillingProject Legs 51, 52 and 53, Holes 417A and 418A. In Donnelly, T.W., Francheteau, J., Bryan, W. B., Robinson, P. T., Flower, M.

F. I., Salisbury, M., et al., Init. Repts. DSDP, 51, 52, 53, Pt. 2:Washington (U.S. Govt. Printing Office), 1185-1199.

Pritchard, R. G., Cann, J. R., and Wood, D. A., 1978. Low-tempera-ture alteration of oceanic basalts, DSDP Leg 49. In Cann, J. R.,Luyendyk, B. P., et al., Init. Repts. DSDP, 49: Washington (U.S.Govt. Printing Office), 709-714.

Rusinov, V. L., Laputina, I. P., Muravitskaja, G. N., Zujagin, B. B.,and Gradusov, B. P., 1979. Clay minerals in basalts from DeepSea Drilling Project Sites 417 and 418. In Donnelly, T. W., Fran-cheteau, J., Bryan, W. B., Robinson, P. T., Flower, M. F. I.,Salisbury, M., et al., Init. Repts. DSDP, 51, 52, 53, Pt. 2: Wash-ington (U.S. Govt. Printing Office), 1265-1271.

Scarfe, C. M., and Smith, D. G. W., 1977. Secondary minerals insome basaltic rocks from DSDP, Leg 37. Can. J. Earth Sci.,14:903-910.

Scott, R. B., and Hajash, A., 1976. Initial submarine alteration of ba-saltic pillow lavas: a microprobe study. Am. J. Sci., 276:480-501.

Velde, B., 1977. Clays and clay minerals in natural and synthetic sys-tems. Developments in Sedimentology, 21: Amsterdam (Elsevier).

563