sign in sign up
<<
Home , Geikielite, Phlogopite, Pyrope

A srr lDY OF 'T'H ~ KI11JfBERLirr1E IN

ELLIO't'rn CCU TY, K~ 1 TJCKY

Frank H. Fowler Geology 570 May 27, 1968 PREFACE

This paper i s pr sented as a fulfillment of the requirement for Geology 570, the Senoir Thesis. This investigation was suggested by my thesis advisor, Dr. c. H. Shultz, who mad possible the actual fi ld study of t he outcrop. The outcrop was visited in December, 1967 and subsequent library and laboratory study was conducted during the Wi nter and Spring quarters 1968. Dr. Shultz also provided thin sections of the rock samples collected at , h site, and provided access to photomicrograph equipment. His advice and introduction to references was also greatly appreciated. Dr . Faure supervised the X- ray i nvestigations.

i i CONTENTS

ABSTRAC'r • • • • • • page 1

INTRODUCTION • 2 Location and History

GENERAL GIDLOGY • 3 l'utroduoticn Stratigraphy Structure

PETROGYDY • • • • • • • • • • • 4 Outcrop Hand Specimen Hea.vy Concentrate Ca.lculat d Chemical Analyses

DISCUSSION • .. • ·• • • • • • • • • • 12 Relation to Regional Occurrences of Si milar Intrusives Comparison with of Siberia and South Africa

Method Q~ Intrusion and Evidence for Formation of Kimberlite Magma and its Condition Upon Intrusion

CONCLUSIONS • •· . . • • • • • • • • • • • 16

REFERENCES CITED • • . • • • • • • • • • • • • 17

iii Figures

Plate, 1, tipre. l • • • Sketch map ct llli.·o'tt OolUt-· lentucq, showing relative position of kim\erlit• (l!Od.iti•d trom Koenig 1'9'6').

Plat.• l; ftll:r 1 2 , • • Kap ·ahow.f.ng outero.pa -<d pattem &f kim'b,erlit.e

Pla.te ! , Pipre l • • • heah eur~aoe o - kimber11ie "har4ba.llk·8 with i1 ·estone and shale inclusion.

Plate . ~ . i'ipre 1 • • • Photomi.cropapb or general s•oti:Qtl or k'imberl:lte sho-S.:". ·. large r.otnlded. oliv:ine, and nall .e\ihe4ral olivine -80 ·,,· . ~ : e,,- opaque ., map.et!t• t and perovl4d. te in of serpent'.in• ani ca~bonate . • • Phc»tomicrograpll ef :perovslcite alteratien rim u-ouad ilmenite, .s.erpentiit:iaat!.on. of oliv.lne, and mapt1-te in gro-UJ!ldmass of serpentin.e and carbonate.

Plat• 4, ·· pre l • • • Photonricro,graph :slloWitlg •BA17 f~actt;tre pattern of' large ilmenite phene,c17sts.

Plate: 4, rtp.r·• 2 ,. • • Photo 1orograJ'h &t kel.,.Phitio reaction rim

around pnet, with .· gnetite' :in se~eatine and carbonate e;n>udmass,.

I.B4u of i•f"ract1o:n ttt Ga.set , Arraage4 tn'\e Groups Acu»rtilt• -..o· Coler. '!'-&ale l • .,, . Chemioal. AJtalyses af lim'berlites •. • • • • • 11

iv The El,lio·'tt County di tremes are basal tic kimberlites occurring a radially arranged groups of apophysis type intrusions centering around large eutc~p ·o.i1 Hamilton Creek. An explos.i ve :fluidization intrusion of a gas.-aolid irlure is indicated by the lack of stnotural d.ef'oJ'fb:tion and contact eta.morphism of the country rock. TThe outcrops have identical mineralogy 'but vary in percent composition of inerals and inclusions because of ditterenoes in the degree o·f alteration, intrusive oonditions, a.nd stratigr&plly ·o.f the country rock. In·clusions of probable upper mantle material, Precambrian basement material, and local •11 rock occur indtH3&ting the probable souroe of the kimberli te as the upper mantle. The rela.tionshi . bet een garnet color and pyrope content is apparently a result of one color being the first melted fraction of the upper mantle While t he other is _the residual garnet caught up by the magma as it i ntruded. Olivine phenoor7sts are of· two ages indicating a pa.use in the intrusive a quence which could, according to Bard t (1965); account for the abseno• of . The 269 m. y . age (Zartman, et al. 19'67) of - .,...._. th~s intrusion corresponds to that of the ultrabasic - of the Kentuelt7-Illinois fluoFspar district. Both lie on one of several deep crtt.eta.1 fa-.ilt zones of the east.em. United St.ates eraton along which intrusion a.ppar.ently took place at various tiJJtes throughout Paleozoic and Mesozoic time.

1 . 2.

INTRODUCTION

The kimb rlites of Elliott County wer the first igneous rocks diseovered in Kentucky. A. R. Crandall of the Kentucky Geological Survey found the intrusives in the valley of Ison Creek around 1881. J . S. Diller, also of the Kentucky Geologic Survey, was told of the outcrop and published sever&l reports of his investigat~ons of it before 1900. Later, Crandall (1910) described this intrusive as a "crater-like dike" with arms radiating north, est, and · southwest from the main body. Diller referred to the rock as a peridotit e until Lewis (1888) coined the term "kimberlite• 1 for the d.iamantiferous plugs of 'the Kimberley area of South Africa to which. these rocks are nearly identical. A kimberlite, as defined by Lewis, is a porphyritic comp0ssed essentially of olivine and its alt eration products. In addition, J1hl ;gopite, ilm.en .te, pyrope, diamond , , and other re present in var,ing amounts. Chemically, kimberlit s are oharact r i z d by lo Si02 and Al203 content, high. gO , fairly high Ti02 and C02, pres ne of P205, and AI2o3 greater than K20 and Na.20 combined. Kimberlite therefore is a notably magnesium-rich ultrabasic rook• The o•terop being studied is located in northeastern Kentuoky in Elliott County. ore :precisely, it occupies the southwest corner of the south est 1/9 of t he Willard geologic quadra.ngl , 83°00' west longitude- 38°07' 30" north latitude (Plate 1, Fig. 1) . Exposures sporadically dot the ridg b tween Ison Creek and Hamilton (Hambleton) Creek for about a mile east of the Ison- Johnson School ( Plate 1 , Fig. 2) . Wi t hin the last few years drilling has be· n done at the largest out crop, labeled "mine" {Plate 1 , Fig. 2) , to determinepossible economic value of t he rook, but although there is a continuing a.ttemp't to promote a mining op ration, nothing of outst anding economic i mportance seems to have been fotind. Recently a magnetic survey was conducted i n an attempt to establi.sh th otual ext nt and form of the i ntntsions. This paper is a petregra io study of this group of intrusives to gain additional infor.rnation concerning the origin a.nd manner of implacement of this kimberlite in pa.rtiottlar and o:f all kimberlites in general. Plate 1 co.

ROWAN

/' I ,..J ,nr' / ,,/ ~ ~ , ,) I ' I I ~ r- '"""'\ , MORGAN CO. ~ , , .. _ ,..r--._ .... r\."-,v-'( LAWRENCE CO. 1------, \ _....,..__ ___ '- I ...... __, LEGEND \ 1 ...... ) I

Figure 1 Sketch map of Elliott County, Kentucky, showing relative position of kimberlite (modified from Koenig 1956).

3~ 0 01'30" --~·~~~~~M~~it);),~~~~~~~~~~-:-- rsoNV.,,,e QtAacl .

Figure 2 Map showing outcrops and pattern of kimberlite in area between Ison Creek and Hamilton Creek.

0 I MILE GENERALE GEOLOGYE

Introduction Elliott County is characterized by steep hills forming even-crested ridges which are separated by narrow alluviated stream valleys. The topography is mature with average elevation between 700 and 1100 feet. Limited farming and grazing is done in the larger valleys, while the streams cutting the ridges are narrow and deep. t

Stratig raphy The region is underlain by a sequence of nearly flat lying Middle Pennsylvanian beds of the upper half of the Breathitt Formation. This consists of a typicalp coal measure sequence predominantly of sandstones, witht interbedded siltstones,t shales, thin coal beds with underclay and limest ones. Sandstone members of this formation cap the hills accounting for the even-crested ridges of the area. The region is a part of the Allegheny Plateau physiographic province •

• Structure Elliott County is located on the west flank of the Parkersburg Synclinorium. Regional dip is to the south and southeast at a very small angle. The dip is reversed locally by faults, small domes and

basins associated with the regional structure.r The intrusives are on the upthrown side of the, Little Sandy Fault, a high angle normal fault extending east-west for about 12 miles, located one mile south of the outcrop area. This fault is part of the more extensive Rough Creek­ Kentucky River fault zone which connects the Elliott County diatremes with the mica-peridotites of the western Kentucky-southern Illinois

fluorspar district. K~Ar and Rb-SrS dating (Zartman,,!i !!.• 1967) shows these two intrusive complexes to be of similar age, around 269 m.y., Late i

Pennsylvanian-Early Permian. TThe EElliott County diatremes show no obvious ·relation to the faults in the area except possibly as pre-existing subsurface zones of weakness along which intrusion took place. Evidence of minor doming or fracturing in the immediate vicinity of the intrusions

is very slight or entirelyt lacking. 4.

PE'l.'ROGRAPHY

Outcrop The kimberlite bodies are best exposed in and on the banks of steep stream valleys. The most obvious exposure is 0. 25 mile south of the Ison- Johnson School. This is a r ough ridge of nhardbank" on the west facing bank of the str eam. On the opposite bank, erosion is less rapid and the outcrop consi sts of "yellow- ground" with residual grains pf garnet and ilmenite scattered through the soi l . Additional outcrops were difficult to locate due to their smal l size, apparently radial distribution, and r esidu l seil cover. Although the rook appears in hand sample to be r elatively resistant, it act ually weathers at about the same rate as the country rook. his inhibits recognition of additional exposures or tracing existing ones topographically. The local extent, identical mi neralogy, and roughly radial outcrop pattern of the& small isolated outcr ops i ndicate that they may be part of the same in· rusi on at depth.

,Hand Specimen Fresh sam les of the kimberlite are dark oli ve green t o nearly black, while weathered surfaces are yellow brown. Yellow of olivine show up prominent ly on specimens of intermediate degrees of

decomposition. It is a holoorystalline, porphyritic ro~k with phenoorysts

of olivine, partial ly serpenti nized olivine, pyropio garnet, ilm~nite , enstat ite, ohrome- , a.nd phlogopite in a serpent ine and carbonate groundmass (Plate 2, Fig. 1) . The size of the phenocrysts varies from

one exposure to another g nerally ranging from 2 ~o 15 mm, but some reach 25 mm across. Anhedral deep- green or yel lo ol ivin s are most comm.on , whi le rounded - black ilmeni te and deep- red, nearly spherical ar more easily seen. En.statite occur s as large anhedr al crystals only at .the "minen ar ea. Phl ogopit 'e occurs spari ngly i n 11 outcrops as subhedrai. crystals, and is particularly ell devel oped at the "mine" out crop. Chrome;di-opside rarely .exc eds 3 mm . It is not common, but stands out because of its vivid green color. Combined , the phenocrysts may compose as much as 60 pereent of a sampl e. Plate 2

Figure 1 FFresh surface of kimberlite "hardbank" with lim stone and shale inclusions. 5•

•, Inclusions are common, . composing 80 percent of one outcrop. The majority are shales of local origin, but limestones and feldspathie nodules are also present. The shale- kimberlite boundary shows large blocks of shale which have been separated from the bedrock by thin outlines of kimberlite. The shale is not strongly metamorphosed as might be ex cted, being slightly baked at most. Limestone inclusions are mor rounded, having traveled further, a.nd show some effects of met somatism. Inclusions of the entire P 1 ozoic section could probably be reoo-gniz d by careful analysis and comparison. Rounded syenite inclusions, possibly from th Precambrian of the upper crust, show

rounding effeo1ts from transport in a more or less solid magma but no. evidence of being altered by it.

Thin Section Thin sections ere examined from several outcrops. '!'he mineralogy of these, excluding inclusions, is identic 1. Variations in the percent of each mineral from section to section mostly reflect differences in the a.mount of Iteration . TThe most obvious feature of the rock is its • high olivine content, ~riginally as high as 80 rcent in some sections before ser entinization. TThus the rook i s a basaltic kimberlite, since i t contains over 50 percent olivine. Lamprop~yrio or ioaoeous kimberlite ~ contains more · than 50 percent phlogo~ite . TTh minerals identified include olivine, ser entine ( variety}, a carbon t e (mostly ) , magnetite, ilmenite, perovskite, ohlorite, garnet, orthopyroxene,

kelyphite, and leueoxen • Although some chrom - diopside ~ , and phlogopite phenocrysts wer identified in hand speeim ns, none could be found in thin section. So e inclusions were found in th se sections. The miner ls can b divided into two large groups. One of primary minerals including phenoerysts of olivine, ilmenite, and garnet with some orthopyroxene, magnetite, and perovskite. The econd group are the alteration products and include carbonate, serpentine, ohlorite, kelyphite, l euooxene, and som pePovskite and magnetite~ The rock is essentially fine-grained c rbon te and serpentine enclosing altered o · vine ~ ph noorysts (Plate 3, Fig. 1) .

6 •

TTh olivine is apparently of two generations. ' he old r is 1- rger, • 5- 12 mm , gen rally anh dral, and greatly • Some grains of this group show undul tory extinction indicating d eform~tion while still under consider ble pressure. Most ar highly serp ntinized, mor so than

crystals of th other groupp . The smal~ r , 0. 5- 3 , ar uhedral and sho little evidence .of having b n st rained or altered. No diff erenc was found in the optical properties of the t o typ s. Both - have 2V of a ppp _roximat ly 90° indicating a oom osition of forst rite90- fay liteio· 'Th ivine lters to antigorite and fin magn tite dust hioh r tain the outline of the form r cryst 1 . . Chlori te s foimd in small quantiti s within th serp ntine in som sections. S rpent ine also exists s a part of the ground.mass being d rived from th pr - xisting

matrix 1 ioh is. now totally altered. A littl serpentine s fo d in limeston inclusions from metasom tic replace ent of min ral grains within ::;th li estone • The inclusions r not greatly -metamorphos d, but d ·e to the high • pr ssure a.nd volatile content of th . original gas- solid magma , most show some videnoe of metasomatio replac ment. 'rhe bound ries of the 'limestone inclusions ar sharp and sho only inor signs of r otion with th groun·dmass. o shale inclusions were found in the sections. A few masses of erpentine and orthopyroxene were observed, hi ch may

rep~ resent alt r ed remnants of up er m&n:tle inclusions. The boundary was ve y obscure and non of th other typical kimberlite minerals were

found inside these ma~ses . Ort ho yroxene, near enstatite in compositi on, also as seen in sever 1 sections a fine ne dl s scattered throughout the matrix. The carbon t , lik th olivine, is ap r ntly of t o types. One is the result 0f late forming carbonate rich solutions and occurs as large singl anhedr 1 crystals several mm ·cross. The other type is more i ntimat ly associated with the serpentine of th groundmass. It occurs s fine gr ined sses, hioh enolos the oth -r min rals of the matrix. Where this carbonate comes into contact with ilmenite, it produces a • reaction rim of fine grained perovski te (Plate 3, Fig. 2) ·• Perovski te also occurs as square brown euhedral crystals up to 0. 01 mm across existing lone in the matrix unassociated with ilmenite. Crystals of this nature are considered primary, or at least formed very early. Ilmenite and magnetite make up ·the opaque minerals. Some ilmenite is in large phenocrysts showing perovskite alteration rims and a sugary texture (Plate 4, Fig. 1). Others occur with magnetite in the matrix as 0.01 mm across anhedral grains occasionally altering to leucoxene. Magnetite in the matrix is usually euhedral , but is otherwise nearly impossible t o destinguish from ilmenite. Garnets range in size from 0. 5-15 mm across. They are nearly spherical and re al ays surrounded by a kelyphitic reaction rim. Kelyphite (Smirnov, 1959) is eompesed of a fibrous inner layer of a.nd amphibole oriented normal to the surface of the garnet with an outer layer .of magnetite, ohlorite and some carbonate (Plate 4, Fig. 2) . This complex mineral a.ssemblege results from the reaction of ga.rn t and t he kimberlite .ma.gma . • On the average, olivine composes "'bout 30 percent of ea.eh section, serpentine about 35 perc nt, carbon te about 25 percent, perovskite 3 percent, magnetite and ilnenite combined about 5 percent, and less than one percent of ea.ch of the other minerals.

Heavy Mineral Concentrate A sample of detrital material was collected from a gully containing

small plaeer~like concentrations of kimberlitic minerals immediately below the "hard.bank" outcrop 0.25 mile south of the Ison-Johnson School . The heavy min rals were concentrated in Bromoform and then screened into different size fractions. No were found eith r in this concentrate or in thin section. Ilmenite , as the most abundant mineral in the concentrate. Thea grains were jet black fragments with conohoidal fracture and

no apparent . any l a~ger grains had a sugary t exture as shown I in Plate 4, Figure 1 ... I I • I I

I P

P e•

Rounded surfaces of the original phenoerysts are pitted and discolored • when pres nt. The pitt$d surface is a result of alteration to leucoxene or perovskite e.nd remnants of these minerals cause the wwhit "

discoloration. A few crystals w re eoated with resulting from the oXidation of magnetit• lamellae within the ilmenite. X-ray diffraction of the ilmenite confirmed the presence of small

quantities of .both ~ematite and magnetite. It was impossible to det rmine the Mg(] content since geikielite, MgO . Ti02, and ilmenite have nearly identieal diffraction patterns and occur in solid- solution with one another. An X-ray scan revealed the presence of several

traoe element eommon to kimber~ites throughout the world. These included , , nickel , oop-pper, zinc, niobium, zirconium, · and rare earths of the Lan.thanum Seri es. Garnets were isolated by magnetic separation and t hen divided

visually into seven groups based on oolor. The index of refraction of about 40 grains in each group were measu~ed and plotted (Table 1) .

The color range from ligh~ brown through red-brown to deep purple was

found to olos~l.Y parallel a rise in n values. TThis reflects an increase • in the content of th garnets. At n=l.74 th eomposition ·is pyrope80-almandine20 and at n=l.75 it is pyrope70-alma.ndine30• This variation will be discusse\d later with regard to the initial agma. formation. The range of measured indicies compares favorably with index data on Siberian kimberlites (Smirnov, 1959) and the James Bay Lowlands kimberlites of· Canada (Skimming, 1960) whose dominant ranges are 1.744- 1.759 and 1. 74-1. 76 respeo\tively. The index of refraction of the g&·rnets was measured within an accuracy of ±.0 . 003. Garnets of color groups #3 , and make up about 80 percent of the t otal. Since garnet color 114 , ·116 out oan b determined by factors other than almandine content, Cao or Cr203 cont ent for instance, th !esults of Table 1 should be considered only as an approximation. his is, in faot, probably the reason for the anomalous values of sample #3 • • I . i •• Table 1 · lndex of Refr ction of Garnet Arranged into Groups According to Color.

'10 <;FIRNE rs (IER COl.Olf c;~OL4P J5

30 JO

10 tO

f 5 15

JO JO

5

0 " I. 76 I. 7S /. 71./ . Index of refraction Garnet Color Group #1 Very light brown •• #2 Light brown, insufficient quantity, not plotted 113 Brown #4 R d-brown 15 Red - ~urple #6 Parple Deep- purpl , nearly opaque #7 p pu p

Enstatite, ohrome-diopside, olivine, and some extraneous sedimentary debri , in decreasing order of abundance, a.ke up the re inder of the heavy mineral conoentrat • Enstatite is a translucent olive green color, which becomes darker and more cloudy with increased alteration. It was most easily separated from chrome-diopside by color since the optic 1 ' properties of both were nearly the same. Chrome-diopside is a lighter transparent vivid green color, which looses some of its luster with alteration. Olivines were simply the cores of serpentinized phenocrysts • th t had eathered out • 10 •

Calculated Che i~al An lys_es • Ch mioal -.nalyses of the v rl.ous outcrops er co puted by substituting kno ·.or aup)Os d in ral eom-posi tion into thin seetion modal analy es. ' h r sult were compar th actual eh ieal ·analys s of various ki b _riites ·(T :bl 2) . od 1 ana.ly s of 1000-2000 c9unts ere made on ch of six sections.

Certain ptions had to be ade co~oerning the chemical fo .ul s of th minerals involved. For olivin , forst rite90- fayalite10 ·s used based on the o tioal prop rty, 2 approximately 90° . For garnet, pyrope78- lmandine22, s used ba ed on the ind x of r fraction. The carbonate -· s assumed t o be .all ca.leite since nothing as known abo:t:tt the dolomite cont nt, if. any. Ilmenite from kimberlite i t ioally rich in gO , which oould no·t be determined.. It was . therefore assum · d to be il Dite as 'that of the Sib rian kimb rlites. Id.' al composition were used for serp ntine, magntite, and perovski t • . o.rthopyrexene and kelyphi t were ~egl _.ct d since n i th r exo ded on · rcent in ! mode, and their co.mpo~itio:p. s . uncertain~ •• The ~nly jor di er pancy betweem the computed: and the actual· chemical a.nalys . otmt of A1 2o3 which ·is. much lo e~ in the computed analy es, ssentially absent. ·Garnet is the only o~_al min "ral . conte.ining ·A1 20; ·and 1 rge garnet pbenocry t ere avoided while &king th odes • . Te count one of th big garnet ould v · produo · · equ · 1~y erroneous high values for 2o3 i"nce they ooeur so sparingly. H20 'l.J:,y. values are lo ··. ,, no serpentine is its only odal sourc , ~ile c~ · cal analy e _also colint unattached water in the rock. 'fr c e:lement are - . naturally ab ent in the · odal an lyses sine they are not a p rt of the formulas of ineral counted in the mode . The overall imilarity of the analys s ould see to or dit th a sumption concerning in ral compo ition. It s lso noted that in spite of differences fro section to section in d gree of alt, ration nd p reent of individu 1 iner ls,' th 8.n lyses ere nearly aiik , showing the over 11 co pasitional quiv 1 nee of separate outcrops. Al o, the sections r bias d, cut to sho specific f atures r ther than r pr senting • _ roe • · Inclusions er voided whil a.king the med s •.. 11 •

Tabl 2 Chemical Analyses of Kimberlit s . • T 1 2 3 4 5 6 -Si02 36. 33 31.60 27 .30 29.e1 27.21 29.21 Ti02 1. 89 2. 02 1.31 . 2 . 20 3.99 2. 79 Al2o3 5.09 ·3.21 3.. 44 2.01 0.22 0. 09 Fe 2o3 7.43 6. 33 . 3.92 5.16 4.52 4 . 20 FeO 3.40 3.37 5.10 4.35 6.15 ·5.12 MnO 0.10 0.30 0. 23 MgO 26 .63· -29 . 45 30 . 30 32. 41 30. 99 33 . 40 Cao 6.78 s.01 12.70 7.69 14.19 12.96 Na20 0.37 0.16 0. 20 0.11 K20 . 2. 43 0. 34 0.26 0 . 20 H20+ 7. 25 11 . 37 4. 63 3. 47 3.59 H20- o.a7 0. 48 C02 1.64 1.96 9. 23 6. 66 8.60 P205 o. 66 0. 94 0.65 0. 35 • TOTAL 100. 00 99 . 99 99. 52 100. 86 100 .00 99. 96 1. Average chemical composition of basaltic kimberlit, , 10 analyses (Nookolds, 1954) . 2. Kimberlite, Kao pipe, Basutoland (Dawson , · 1962 , p. 551}; M. H. Kerr, analyst. 3. Elliott County,- Kentucky, kimb rlite (Koenig, 1956 , p. 49); M. E. Coller, analyst. 4. Elliott County, Kentucky, kimberlite (Dill r, 1887, P·· 25);

TT. • Chat a.rd, analyst. Ineludes 0. 43. Cr203 , 0.05 Ni , a.nd o.2a.so3 in total but not in the table. 5. Average of six analyses based on modal analyses of thin sections. Elliott County, K ntucky. 6~ Calculated analysis from modal analysis of thin section I P67KY6A . •• Elliott County, K ntuoky• 12 •

• DISCUSSION Kimberlit s are usually found around the margine of st ble conti nental eratons. The eratons are not entirely st tic, ho ver,

particularly near the margins where they may contain large sc~le crustal deformation structures • . The Elliott County kimbe:dite lies along .one such feature, the Rough Cr ek-Kentucky River fault zone, which extends westward into issouri ~ The ·mica- peridotites of the western Kentucky­ southern Illinois f luorspar district are also located along this feature.

Crusta.1. deformat ion apparently created this zon~ of eakness along which these ultrabasics were then intruded. Zartman, !!, al. (1967) have determined th K-Ar and Rb- Sr ages of both intrusiv s from phlogopite phenoorysts. The results of both types of age determin tion were overlaping and averaged 269 m. y. , Early Permia.n; for both areas.. This

indicates that both ultrabasic compiexe~ are related to the time of formati on of .the major zone o~ weakness. The composition difference between the t o areas is probably a factor of the depth of formation of the original ·magm and the intrusive history of each. • Many other a.lkalic intrusive eompl xes oeeur along the e stern craton of North America. · Each seems to b rel t d to a similar s tructur~ feature, but the structural features are apparently not r elated to one another. Int rusions. occurred spora.dioa.lly throughout most of Paleozoic and ·esozoic tim reflecting a history of repeated stress, of unknown ca.use; agai nst the sta.bl craton • .This relationship is also true of other continents, and it is interesting to note th t the produc d is nearly identical. Con1pared with the Elliott County exposure nearly identical descriptions w.ere given for Siberian and South African kimberlites . 'f'h minerals present, their relati ve proportions and the ch mical compositions are very close ( abl 2) . Inclusions are also characteristic. However, in other kimberlites inclusions of au posed up. er mantle origin seem to be more common • • A characteristic of many kimberlites not observed in the Elliott County • outcrop is the presence of inclusions from higher strati graphic levels. Hearn (1968) describes identifiable ki~berl it e inclusio~ s that represent formations more than 4000 feet stratigraphieally higher than the unit adjacent to the kimberlite intrusive at the present erosion surface. The presence of magnesium minerals like perovskite and magnesium­ rich ilmenite are characteristic of kimberlites everywhere, and along with low silica content, distinguishes kimberlit es from most other igneous rocks. The carbonate is part of the original magma. It has been suggested that the carbonate could be derived from assimilation of carbonate inclusions, but kimberlites of shield areas such asc.tbose in Canada have not penetrated carbonate formations and yet contain the normal amount of carbonate in the matrix. The kimberlites of Siberia, South Africa e.nd elsewhere occur as pipe-like intrusions similar to the Elliott County exposure. They may be larger, as in South Africa, but they still show the same relat ionships t o the enclosing country rook. Meta.morphism is very slight if recogni:Za.ble a.t all. The type of intrusion seems to be nearly the • same i n all localities. Judging from their worldwide similarities of mineralogy, texture and structure, it seems likely that they all were produced from a homogeneous source. This, i n addition to the presence of occasional eclogite inclusions and other typically high pressure mineral phases like pyropic garnet a.nd diamond indicate that their source is somewhere within the upper mantle. Kimberlite intrusion apparently results from crustal movement producing zones of structural weakness in the lower crust and plastic deformation of the upper mantle. Any rook will flow if subject to enough pressure. Because of the tremendous pressure at that depth, the upper mantle becomes mobile without necessarily melting. The high density "magma" intrudes the crust along the zone of weakness with a relatively high velocity, working its way upward toward the surface. At the lower confining pressures of the upper cmlst the gaseous

components expand fluidizing the ma~atic mass, which breaks its way 14 •

by stoping through overlying rocks while producing only minor structural • deformation in the immediate. vicinity. At the surface, local joint systems, fault . zones or lithologio units. will contr ol th actual outcrop patt ern. The percent of inclusions ould d pend upon the gaseous content of th particular "magma", the lithology intruded, th . g ometrio relationship of th particular pipe to the main bedy and on ho ell the gas pressure was maintained .throughout intrusion. As this gas- solid "magmau rises it picks up pieces of country rock fr9m all depths. These inclusions are rounded by abrasion, the degree of rounding being a function of their resistance to abrasion as ell as the distances t r aveled fro their sources. Th high pr essure and gas content cause significant meta omat ic replacement withi n some .inclusions. , however, is nil since the "magma'• is i ntruded relatively cold, less than 500°C. No flow structure or preferred orientat ion of crystals is present . According to Mikheyenko and ena.shev {1962) kimberlite b gins to melt at 1230° and melts completely at 1500° .· If kimberlites ere ver molten, much greater contact meta.morphism would be e.xpeot d. Because of this, all phenocrysts must have formed very

1 • early , being pres rved by maintaining a relati~ ly high pressure unti l shortly Befor movement ceased. Any pause in the intrusion cycle would l low pressur to escape and the 1 ss stabl high pressure miner ls would b n to break down. According to Bardet (1965) this could account for the l ack of diamonds in the Elliott County kimberlites. The presence of two gener tions of olivine woald further indicate a break in th intrusive cycle at some level. The existance of a r nge in composition in garnet indicates T dif ferent iation or part ial melti ng at some point in the ki mberlite history. Since it is vident that the kimberl ite was intruded as a solid, di ffer ntiation by partial melting woul d have to have been from a previous heating event. Another possibility would be ion diffusion within the mantle creating a compositi onal gradati on with depth. In this c se the color and compositi on series would indicate a certain • depth of format ion of the kimberlite ''magma .. . 15.

However formed, the garnets were rounded b~ abrasion and developed their kelyphit io rim due to reaction with the gaseous "magma" during the final stage of intrusion as movement ceased. Serpentinization would begin as soon as the relatively cold ttmagma" got below 400- 5000 near the surface. Carbonitization is a late stage effect of the gaseous fraction. Carbonate, from ground water, is also present as crack and fracture fillings in the kimberlite •

•

• 16 •

• CONCLUSIONS 1. The Elliott County, Kentucky, diatremes are basaltic kimberlites strongly resembling kimberlites of Siberi , South Africa and elsewhere in min ralogy, ohemieal composition, high magnesium content, and texture.

2. It is an a .ophysis type intrusion surrounding a single pipe at depth. The mineralogy of each branch' pipe is identical while varying in amount of alteration and inclusion content.

3. The souroe of the kimherlite is the upper mantle and therefore kimberlite intrusions potentially could contain inclusions of all intermediate formations. Garnets form a parallel color and composition series th t either represents partial melting of one type of garnet and the inclusion of the other as r esidual matter, or indicate the depth within the upper mantle from which the kimberlite nmagma" as derived b cause of a compositional gradient • due to ion diffusion. 4. The kimb rlite was intruded as a relatively cold gas- solid crystal mush by fluidization. The maximum age of the intrusion is around 269 m.y., Early Permian. A two stage intrusion or m jor pre sur loss at some point is indicated by the two generations of olivine phenocrysts and the absence of diamond.

The Elliott County kimberli te is relat fi'.0t , the regional structure, as are th other lkalic intrusives of eastern. North America.

Ho w ver, th elements of the r giona.l structur , namely the various fault zones, along which the intrusions occur, are not relat d in time. Intrusions oeeurr d sporadically throughout much of Pal ozoie and Mesozoic time along the astern edg of the North Americ • er ton • REFERENCES CITED

1. Bardet, M. G., 1965, Diamantiferous dia.tremes: Eeon. Geol., vol. 60 , no . 4, p. 827-829.

2. Crandall, A ~ R., 1910, Coals of the Licking Valley Region and some contiguous terri tor y, i ncluding also an account of Elliott County and its dik s: Kentucky Geol . Survey Bull. 10, p. 70-90 . 3. Dawson, J . B., 1962, Basutoland kimberlites: Geol . Soc . Amer. Bull., vol. 73, p. 545-560 . 4. Diller, J. S., 1887, P ridotite of Elliott County, Kentucky: U. s. Geol . Survey Bull. 38 , 3lp. 5. Hearn, B. C. , 19§8, Diatremes with ki mberlitie affi nities in North­ C.ntral Montana: Sci nee, vol. 159, no . 3815, p. 622- 624 . 6. Koenig, J . B., 1956, Th of certain ign ous dikes of Kentucky: Kentucky Geol . Survey, ser. 9, Bull. 21 , 57P• 7. Lewis, c., 1888, The matrix of the diamond: Geol . ·ag., 5, p. 129- 131. • 8. Mikheyenko, V. I ., and Nenashev, N. I ., 1962, Absolute ge of formation and relative age of i ntrusion of the kimber lites of Yakutia: Int,- Geol . Rev,, vol. 4, no. a, p. 916- 924. 9. Nockolds, s. R., 1954, Aver age chemical composition of some i gneous rocks: Bull. Geo! . Soc. Amer., vol. 65 , p. 1007-1032. 10. Ski mming, T., 1960 , Unpublished report on kimberlite indi cator minerals, James Bay Lowland: Seloo Expl. Co . Ltd., Tor9nto. 11. Smirnov, G.· I ., 1959, iner alogy of Siberian ki mberli tes: Int. Geo! . Rev., vol. 1, no . 12, p. 21 - 39 . 12. Zartman, R. E., Brook , M. R. , Heyl , A. V., and Thomas , H. H., 1967, K- Ar and Rb- Sr ages of some alkalic intrusive rooks from central and eastern United Stats: Am . Jour. Sci. , vol. 256 , no . 10, p. 848- 867 • •