Composition and provenance of the Lilljeborgfjellet Conglomerate Haakon VII Land, Spitsbergen TORE GJELSVIK Gjelsvik, T . 1991: Composition and provenance of the L.illjehorgfjellet Conglomerate Haakon V11 Land, Spitshergcn. Polar Resenrch 9(?). 141-154. The Siktefjellet Group of late Silurian or early Devonian age. consisting of the Ldljehorgfjellet Con- glomerate and thc overlying Siktefjellet Sandstone, is generally accepted as thc oldest part of the Old Kcd Sandstone in Spitsbcrgcn. Most of the ckrsts of the conglomerate arc only slightly roundcd a n d conaist oi htholopies typical for the underlying basement. A minoi component of quartz porphyry clast, is prescnt: these are well-rounded. indicating a longer transport. The provenance of thc quartz porphyry cl;i\ts is discussed in relation to the known outcrops of quartz porphyry in Svalbard. one occurring in the ncigh- bourhood of the conglomerate. thc other one5 far away. The quartz porphyry in close proximity i \ younger than the Lilljchorgfjrllet Conglomcratc and therefore no! a possible source rock. A closc petrographic and geochemical comparisoii with the quartz porphyries at three localitics in Nordaustlandet (15G-200 kin in castcrly dircctioii and ot prohahle Grenvillian a g c ) is presented showing many aimilaritics. hut enough differences t o qucstion their interrclationship. Thc porphvrics of the Hornsund area (300 kni in routhcil! direction and also of probable Grenvillian age) are found to hc chemically and petrographically distinctly different from the Lilljeborgfjellct clast porphyry. Metai-hyolice rcported from the Plaiicrficllii a n d Hitr- kerbreen Groups in Ny Fricsland arc not comparahle with thc cliist porphyry. As n o unquestirinahlc w u r w rock among the quartz porphyries is known in outcrop. the possibility of a hidden or complctel! crodcd parent rock ia considered. Tore Glelsuik. Norsk Polnrimtinttt, P . O . Box 158. N-1330 Oslo Lufthaon. Norwuccy. -Anah Introduction Bay conglomerate to the red and yellow marble During a sledge expedition in 190Y, 0. Holtedahl (1026) investigated the Devonian Red Bay series around Raudfjorden. T h e series which had been reported previously by Nordenskiold (1892) is confined by north-south trending faults and con- sists of a basal conglomerate and overlying sand- stones of typical Old R e d Sandstone (ORS) affinity. In Rabotdalen, on the east side of the fjord, Holtedahl noticed two kinds of con- glomerate, o n e coloured red and yellow and dom- inated by marble and dolomite clasts. the other more polymictic boulder breccia of grey colour. Although the conglomerates were separated by a distinct fault, Holtedahl considered them t o be only different facies of the basal Devonian con- glomerate. later named by Friend (1961) the Red Bay conglomerate. G e e 8 Moody-Stuart (1966) conglomerate and the overlying red brown quartz chip conglomerate of earlier published studies. When investigating the Lilljeborgfjellet Con- glomerate, the present author noted a striking difference between most of the clasts. which are only slightly rounded and consist of lithologies common in the neighbouring basement: a few are well-rounded and composed of quartz porphyry. The latter are unknown in the basement except for some occurrences in Nordaustlandet and the Hornsund area. These observations raise the question of a very long transport distance of the quartz porphyry clasts versus the possibility of hidden or eroded occurrences of quartz porphyry at a closer distance. Extent and lithology found that the grey breccia, which they named the Lilljeborgfjellet Conglomerate. is the basal - . unit of a conglomerate-sandstone succession Except for some areas which are covered by gla- resting unconformably on metamorphic rocks and ciers, the 150 t o 400 m thick Lilljeborgfjellet Con- discordantly overlain by the R e d Bay conglom- glomerate extends as a continuous unit from the erate. They proposed the name Siktefjellet G r o u p Rabotdalen pass east of Raudfjorden to the for this succession, restricting the use of the R e d Siktefjellet mountain o n the northern share of 142 T . Gjelsvik Fig. I . Geological sketch map. extracted from Gee & Moody-Stuart (1966. fig. 1). Locations of Nordaustlandet quartz porphyries: B = Botniahalvoy. S = Sabinebukta. POL = Prins Oscars Land (Rijpdalen). Location north of Hornsund: V = Vimsodden. h'F = Ny Fricsland Composition and provenance of the Lilljeborgfjellet Conglomerate 143 Liefdefjorden (Fig. l), a distance of 14 km. It generally dips 3&50”WSW. In Puddingen, to the north of Rabotdalen, an isolated outlier of “a red quartz conglomerate” was found by Holtedahl (1926, p. 11) who considered that it belonged t o the Red Bay conglomerate series. However, Gee & Moody-Stuart (1966. p. 62) included it in the Siktefjellet Group. This author has found two additional, isolated occurrences of the same con- glomerate on either side of the Hannabreen gla- cier tongue in Liefdefjorden. Gee & Moody- Stuart (1966) also considered the possibility that the marble conglomerate at the promontory of Konglomeratodden on the western side of Raudfjorden might be a separate unit of the Siktefjellet Group, without giving conclusive evi- dence. At its northern end i n Rabotdalen (Fig. 1). the Lilljeborgfjellet Conglomerate tapers out with a basal breccia resting on an irregular basement of schists and gneisses. When visible, the contact is mostly sharp, in part tectonically disturbed by steeply west-dipping shear planes. In other places, the immediately underlying basement is coloured red by a network of small, hematite-bearing veins down to a depth of 20-30m below the uncon- formity. Also, the basal part of the conglomerate comprises numerous clasts of gneiss containing the same hematite vein system. In one place Hol- tedahl (1926. p. 12) also noted a transitional layer of huge, angular pieces of crystalline rocks between the conglomerate and the basement rocks. The breccia in Rabotdalen consists of very coarse (up to 1 m), angular clasts of gneiss, marble, and metabasites (Fig. 2), many of which are garnet-bearing and of medium t o high meta- morphic grade. Similar lithologies occur in small areas of the adjacent crystalline rocks (Gee 1966, chap. 6). Both stratigraphically upwards and lat- erally southwards the conglomerate clasts gen- erally become more rounded, the proportion of quartzite increases, and especially in the tran- sition zone to the overlying Siktefjellet Sandstone the clasts diminish to form a pebble conglomerate interbedded with arkosic sandstone. In the inter- mediate and lower levels the clasts are still mostly massive, and particularly in the basal part angular clasts of crystalline schists or gneisses are ubiqui- tous. Clasts of coarse-grained marble are also common, but dolomite marble, which is so fre- quent i n the Red Bay conglomerate, is very rare. Less common. but nearly always present, are the well-rounded clasts of quartz porphyry. These are much better rounded than the clasts of other lithologies (Fig. 3). In most cases they make up only 1% or less of the clasts. In one place, however, as much as 5 1 0 % has been observed. The size ranges from a few cm to 1.5 m, but most clasts are small. The subordinate matrix of the Lilljeborgfjellet Conglomerate consists mostly of small pieces of the same rock types as in the larger clasts, perhaps with a greater proportion of mica- rich schists and gneisses. Unfortunately. this study lacked the necessary logistic support for a detailed sedimentologic Fig. 2 . Angular clast of augcn gneiss (below center) in massive Lilljeborgfjellet Conglomerate, Rabotdalen. 144 T . Gjelsuik Fig. 3. Wcll-rounded clnst of quartz porphyry. 1111 long, in Lilljehorgfjellet Conglomerate. eastern hordcr ot the Snout of Hannahrecn. investigation of the Lilljeborgfjellet Conglom- erate trcvii Rabotclalen all the way to Siktefjellet. Most o f the observations ;ire from exposures in the vicinity o f Raudfjorden in the north and Lief- defjorden in the south. T h e conglomerate is t o a great extent c w e r e d bx debris or snow: the best exposures are either inaccessible o r too hazardous for field observations. T h w . i t w i i s not possible to investigate n h e t h e r there are any internal stratigraphic relation\hips defined by the quartz porphyry clasts. N o r was i t possible to measure an! lateral variation in their frequency distri- but 1011 . The lithology and angular to subrounded shape of the clasts. except for the small number con- sisting of quartz porphyry and perhaps for the most rounded quartzites. indicate a short trans- port of the components of the Lilljeborgfjellet Conglomerate. All lithologies. again with the exception of the quartz porphyry, are comparable with those of the underlying o r adjacent meta- morphic rocks (Holtedahl 1926; Harland 1961; Gee & Moody-Stuart 1Y66). It should be especially noted that exposures of garnet-bearing. medium to high-grade metabasites in Spitsbergen are restricted to the adjacent Richarddalen Corn- plex (see map. tig. 3 in Peucat et al. 1989). O n the other hand. the quartz porphyry clasts present an enigma. at least a contradiction. Their uell- rounded shape. despite the tough and hard com- position. testifies of a longer transport than for the other clasts. However. quartz porphyry has nowhere been observed in the Hecla Hoek suc- cessions of northwestern Spitsbergen. Exposures of quartz porphyries in Svalbard Quartz porphyry outcrops have been reported from five areas in Svalbard. T h e o n e nearest to the Lilljeborgfjellet Conglomerate is located on the Frxnkelryggen ridge in Raudfjorden. o n l y about 2-3 km southwest of Lilljeborgfjellet. Three other outcrops are located 150-200 k m to the east. on the island of Nordaustlandet (Fig. 1). Quartz porphyry has also recently been found on t h e western coast of Spitsbergen, a little to the north of Hornsund. Fr:rtrrrkelrj,ggeti u r t x In the Frmkelryggen a r e a . two beds of "welded t u f f " . each 4 ni thick. are interbedded in the F r z n - kelryggen Sandstone which is situated in the middle of the Red Bay G r o u p (MuraSov et 31. 1983). Besides phenocrysts of quartz and feldspar i n n microfelsitic groundmass. the welded tuff contains about 30% clastic fragments and might rather be called a tuffite. In this way i t differs from the quartz porphyry clasts and the Nordaustlandet quartz porphyries, none of which contain extraneous clastic components. Plagioclase phenocrysts make up 3 - 5 Y of the rock, ortho- clase 1-29>, The sizes are (1.2-0.4 nim and 0. I - 0 . 2 mm respectively. Orthoclase phenocrysts are completely altered t o clay minerals. The groundmass is excessively replaced by carbonate. mica and chlorite. T h e rock is mottled with a carbonate mineral. a phase which also occurs in Composition and provenance of the Lilljeborgfjellet Conglomerate 145 Tuble 1. Chemical composition (anhydrous) of quartz porphyry from Fraenkelryggen and from the Vimsodden conglomerate, and of posttectonic granite from northwestern Spitsbergen. N o SiOz Ti02 A1203 Fe,03 FcO MnO MgO CaO NazO KZO P,O, Total 1 71.15 0.18 17.27 1.32 0.79 0.11 1.51 1.85 1.60 4.06 0.13 99.97 2 76.69 0.22 14.00 1.16 0.71 0.06 1.20 1.93 1.45 2.50 0.13 100.00 3 73.91 0.20 15.64 1.24 0.75 0.09 1.36 1.89 1.53 3.28 0.13 99.99 4 70.01 0.22 12.73 4.19 0.65 0.02 0.60 1.59 0.23 9.59 0.11 99.99 5 62.22 0.18 14.67 2.72 1.60 0.41 1.31 5.45 0.30 10.97 0.16 100.00 6 66.12 0.20 13.70 3.46 1.13 0.22 0.96 3.52 0.27 10.28 0.14 100.00 7 70.94 0.31 15.07 0.49 2.06 0.04 0.59 1.87 3.15 4.44 0.07 99 83 No. 1 Welded tuff, Fraenkelryggen, upper layer. No. 2 Welded tuff, Fraenkelryggen. lower layer. No. 3 Average of 1 and 2 No. 4 Porphyric clast, Vimsodden conglomerate, Hornsund. No. 5 Aphanitic clast, Vimsodden conglomerate, Hornsund. No. 6 Average of 4 and 5. No. 7 Average of 8 samples of granite dikes, northwestern Spitsbergen (Hjellc in Hjelle & Ohta 1974, table 3) numerous microfissures. The upper bed has a 10 cm thick sole with quartz and orthoclase phenocrysts up to 5 m m in length. Two major element chemical analyses, one from each bed (MuraSov et al. 1983, p. 98), are recalculated to 100% (anhydrous) and presented as no. 1 and 2 i n Table 1. Hornsund area Smulikowski (1968, pp. 108-111) reported the presence of slightly metamorphosed rhyolite clasts in a Proterozoic, polymict conglomerate at Vimsodden (V in Fig. 1). According to his description, the clasts only contain phenocrysts of potash feldspar, maximum 2 x 5 mm in size, imbedded in an aphanitic, dark grey matrix and usually exhibiting preferred orientation con- cordant with the elongation of the pebbles. The matrix consists mainly of quartz and microcline with minor amounts of secondary biotite, sericite and chlorite. Some of the phenocrysts are crushed and deformed, resulting in the crystallization of quartz, biotite and carbonate in internal fissures and along the grain boundaries. The chemical compositions of one porphyric and one aphanitic rhyolite pebble (Smulikowski 1968, table 2), recalculated to an anhydrous composition, are reproduced in Table 1 (nos. 4 and 5 ) . The rhyolite clasts described by Smulikowski do not contain phenocrysts of quartz. However, according to Y . Ohta (pers. comm.) both rhyolite and quartz porphyry lenses have been found in the adjacent Skslfjell amphibolite complex (Birkenmajer & Narebski 1960, fig. 1) which is probably of Gren- villian age (Peucat pers. comm. 1991). Like the rhyolite pebbles of the Vimsodden conglomerate, the chemical composition of these acid vulcanites is characterized by extremely high potash and very low sodium contents. Nordaustlandet areas Quartz porphyries occurring on Nordaustlandet (Fig. 1) appear in the form of massive bodies, plugs and dikes in Botniahalvoya (B) and Sab- inebukta (S), whereas in Prins Oscars Land (POL) they are strongly foliated and appear as folded sheets. The massive rocks are only slightly altered, mainly by sericitization of the plagioclase phenocrysts and the groundmass, whereas in the POL sheets intense sericitization and carbonate replacement have often led to a complete destruc- tion of the feldspar phenocrysts (Flood et al. 1969; Ohta 1984). However, the metamorphic processes have not obliterated the typical por- phyric texture of the rocks nor the characteristic late magmatic corrosion features of the quartz phenocrysts. The Botniahalvoya quartz porphyry has been dated by the Rb/Sr whole rock isochron method t o 766 Ma by Gorochov et al. (1977), and has been recalculated by the same authors to a model age of 970 Ma. The quartz porphyry clasts in the Lilljeborgfjellet Conglomerate resemble the Nordaustlandet rocks, both in the frequency and in the composition of the various phenocrysts (Table 2 ) . The feldspar phenocrysts have not been separated in Table 2 due to the complex 146 T . Gjelsvik Tubk Z Proportions o f matrix and phenocrystc. Location n Matrix Phenocrysts q u a r t z Feldspar ~- Haakon V I I Land (LC' cla\ta) 13 55-81 3-26 12-25 Botniahalvo)a ( B ) 6 5 S l . i . h i 3 11-25 Sahinobukts ( S ) 2. hJ-67 1&13 Zk-2.1 Prim O5cars Land ( P O L ) 7 Xo-89 h ( & I I Total Nordaustlandcr I 1 59-89 3-13 G 2 5 The pcrccntagc\ i l o m Xordaustlandet are calculatcd from Ohta ( I Y X 4 . tahle 2 ~ ) . intergrowth of the alkali feldspars in the clasts. to be described below. T h e apparent lithologic similarity between quartz porphyry clasts of the Lilljeborgfjellet Conglomerate and the Nord- austlandet quartz porphyries prompted a detailed petrographic and geochemical study of the former as a basis for comparison with the relevant data of the latter (Flood e t al. 1969: O h t a 1984). Petrography In both the Lilljeborgfjellet and the Nordaustlan- det suites. the size of the quartz phenocrysts is mostly between 2 and 5 mm: the feldspars a r e a little larger. However. in o n e of the clast samples the feldspar phenocrysts measure up to 2.5 cm. In the massive quartz porphyries of Nord- austlandet the alkali feldspar phenocrysts are composed of homogeneous orthoclase or finely lamellar perthite (Fig. 4). In the Lilljeborgfjellet Conglomerate (LC) porphyry clasts, the alkali feldspar phenocrysts display a coarse, patchy intergrowth of potash feldspar and sodic plagio- clase (Fig. 5 ) . T h e uniform extinction of t h e pot- ash feldspar domains in the plagioclase host crystal indicates that t h e pattern is formed by exsolution rather than by replacement. Some- times the potash feldspar forms an outer rim or follows the cleavage directions in the host crystal. In both cases the plagioclase phenocrysts a r e com- posed of oligoclase which in Nordaustlandet is evenly and densely clouded by sericite. In the LC porphyry clasts they a r e replaced by more coarse- grained sericite in irregular patches, along veins or in the marginal zones of the grains. T h e matrix of the Nordaustlandet quartz por- phyries is microcrystalline and mostly granular, evidently d u e t o recrystallization. Equigranular quartz and feldspar crystals. usually less than 0.02 m m , mixed with scaly sericite. are the main constituents. In a few cases a primary texture of platy. radial feldspar in a spherulitic pattern is Fig. 4. Quartz porphyry. Botniahalv~ya, Nordaustlandet. Fig. 5 . Sericitized, foliated quartz porphyry, POL (Rijpdalen). Phenocrysts: quartz. plagioclase (rhomboidal), perthite (lower Phenocrysts: quartz. (left center, diamond shaped), sericitized left, extinct). 32x potassium feldspar (right center). 1 3 X . Composition and provenance of the Lilljeborgfjellet Conglomerate 147 Fig. 6. Rounded phenocryst with intergrowth of plagioclase (white) and potassium feldspar (dark) in clast porphyry from Lilljeborgfjellet. Note jagged border of the phenocryst d u e to recrystallization of the groundmass. 13x. observed, indicating devitrification. In the POL rocks. the sericite is sometimes slightly coarser and concentrated in densely foliated bands (Fig. 6). The contact of the phenocrysts appears mostly sharp and smooth, although under high mag- nification it can be seen that the sericite crystals intersect the margins. Stronger recrystallization of quartz, chlorite, muscovite, and sometimes calcite is seen in strain shadows of the pheno- crysts. In the L C porphyry clasts recrystallization of the matrix is more advanced than in the Nord- austlandet porphyries (except perhaps, for some in POL). A granular texture dominates. the grain size of the groundmass is usually in the range of 0.04-0.1 mm, i.e. twice as coarse. T h e matrix minerals are the same as in Nordaustlandet: quartz, feldspar and sericite. Potash feldspar F g . 7. Glomeroporphyric feldspar with opaque seam along the margin. Clast porphyry, Lill~eborgfjellet. 1 3 ~ . seems t o b e the most abundant feldspar, making up perhaps 30% of the matrix. T h e phenocrysts are strongly indented by the matrix minerals, including chlorite and sometimes calcite. Iron ore is present in clusters of large, often idiomorphic crystals together with chlorite and calcite in cross- cutting veins and as tiny crystals in the matrix. In one of the clast samples. some plagioclase phenocrysts are encrusted by an opaque rim indi- cating quick, subaerial cooling. They also have a peculiar glomerophyric texture not observed in the other samples (Fig. 7). A t the same time the matrix often shows a spherulitic texture. T h e source rock of this clast would be of pyroclastic origin. Both in the porphyry clasts and in the Nordaustlandet porphyries, clusters of coarse crystals of quartz, calcite. muscovite, sphene. and iron ore are seen. perhaps representing more o r less recrystallized amygdales. Chlorite, muscovite and iron o r e also occur, derived through replace- ment of larger biotite crystals. To sum up the petrographic evidence, the quartz porphyry of the L C clasts porphyry is in most respects similar to those in Nordaustlandet. The main difference is found in the structure of the alkali phenocrysts and the generally more advanced recrystallization of the former. T h e two associations both comprise subvolcanic and pyro- clastic units which have been metamorphosed in greenschist facies. Geochemistry The major element composition of the L C por- phyry clasts is given in Table 3, which also pro- vides the average ( A v . ) compositions of the various oxides. According to the T A S classifi- cation all the analysed rocks a r e high-K rhyolite (Le Maitre 1989, figs. B14 and 15), although a few a r e close to the borderlines of dacite and trachyte, respectively. T h e average compositions of the Nordaustlandet porphyries a r e included in the table for comparison. There is good agree- ment for most of the elements, the main dif- ferences being lower Al and P , much higher Fe3+, and a little higher C a in the L C clast porphyry. The SiOz content of the clast porphyry agrees with the porphyry of B o t n i a h a l v ~ y a . However, the latter has an even spread of SiOz between 70 and 7 5 % , whereas the clast porphyry has two distinct concentrations of 70 and 74%. K is nearly the same as in the B o t n i a h a l v ~ y a area, but dis 148 T . Gjelsuik Table .T, Major clement composition (anh!drous bascl of quartz porphvq ciast, in Li~ljehorgfjellet Conglomerate. TIO: AIIO. Fc:Ol FeO M n O MgO CaO Na,O K 2 0 P 2 0 1 Total 78-GJ. I I 1C 78-Gj. I 5 1 A 78-GI I 5 1 8 7X-Gj.I5?A 78-Gi 1578 78-GJ 157C 78-GI l5Y.A 78-GJ. 15qB 78-Gj. 15YE 78-Gj.177D 7X-Gj.182 78-Gj. I Y 1 U Av. (I31 78-GI. 1 5 Y ( ~ 73.7s 73.6’) 73.24 70.22 70.115 70.19 711.38 74.0’) 71.53 74. I I 73.51 73.41 71 12 72.3.7 (1.37 U.48 0.13 (1.67 (1.63 0 78 0 72 0 32 0.5X 0 4 1 0.43 0.66 0 31 (1.57 A \ . B ( 7 ) 72.77 0.3(1 A\.. S(3) 73.82 0 . 1 9 A \ . POL(%) 75.13 0 3 2 I2 83 12.x3 12.87 13.26 11.83 13 17 12.93 13.10 11.84 12.7X 12.89 1 l . 2 0 12.74 11.81 2.14 0.46 1.92 1.13 1.57 1 . 1 1 3.61 1.31 4.59 1.15 2.16 3 . m 4.1 1 1.08 0.55 1 3 4 3.73 0.78 0 . 4 3 2.05 -._ 7 i 6 0.28 1.64 1.75 1.81 0 47 2.37 1.66 15.06 0.57 1.44 14.81 0.37 1 5 1 14.54 o.no 1.6s 11.01 0 112 O.I)2 0.03 0.03 0.07 0.02 0 (13 0.02 0.03 u.02 0.m 1).03 I) 03 I1 02 I).O? u.03 ~ 0.62 0.72 1.09 1.35 0.44 1.37 1.37 1.73 0.63 0.50 0.27 1.60 0 . 3 3 n.92 ~ 0 . 5 1 u 79 11.81 1.67 0.94 0.31 1.49 1.50 1.36 1.69 2 67 0.95 0 91 1.70 5.77 1 .63 0.78 2.16 6 . 5 2 1.65 7.65 3.40 4.50 3.49 4.11 2.80 4.73 1.96 5.43 3.30 5.25 3.23 4.84 2.86 5.78 2 . 8 4 5.82 3.32 5.06 3.07 4.04 1.53 6.39 2.75 5.31 0.04 0.06 0.06 0.13 0.14 0. I5 0.14 0.04 0.12 0.05 0.06 0.12 0.04 0.09 100.00 Y9.YY 1ou.00 YY.YS 99.99 100.01 99.97 Y9.97 YY.99 99.99 YY.YY 99.90 1 ou 011 1.09 7 . X 1 5.16 0.23 0.78 3.14 3.Y4 0.19 1.06 1.49 3.91 0.11 Sdmple n o 78-GI. 1 I IC is collected near the Hogeloftet ridge north of Liefdcfjorden. The other samples arc from the area hetwccn Kahotdalen and Lilljchorgfjcllct on the eastcrn 51de of Kaudfjordcn. tinctly higher than in the two other areas in Nor- daustlandet. In Figs. %11 t h e distribution of the elements of the LC clast porphyry is shown by crosses in the various oxide variation diagrams given by O h t a (1984. pp. 85-88, 90) for the Nor- daustlandet rocks. Fig. 8 A shows that the total alkali content of t h e L C clast porphyry is in the same range, or slightly higher range, as that of Botniahalvcbya. In the alkali variation diagrams (Fig. 9C) it is seen that especially the K ? O content of some samples is higher. although most of the samples are within the same range as the Bot- n i a h a l v ~ y a rocks. Since the L C clast porphyry has been subjected to extensive sericitization of the matrix as well as the feldspar phenocrysts, it is conceivable that some K has been introduced. However, the Al diagram (Fig. 9 C ) shows that the A1203 content of the L C porphyry clasts is consistently lower than the content of the clasts of Nordaustlandet. Moreover. the A1 content of the former lies in the range of representative rhyolites of the andesite-rhyolite family (Car- michael et al. 1974. table 2-2). These relation- ships suggest that the source rock of the clast porphvry initially has a high amount of potash feldspar, rather than being subjected t o potash metasomatism at a later stage. T h e relevant dia- grams in Fig. 9C demonstrate that the trends of FeO, MgO and C a O in t h e clast porphyry are the same as in the Nordaustlandet rocks. However, both Fe and C a have been mobilized and perhaps introduced during the recrystallization of the clast porphyry. as shown by the vein-like or cleavage- related distribution of iron oxides in some samples, and by the numerous cross-cutting cal- cite veins. Because of the latter, o n e very high Ca figure was discarded from the diagram. According to the data for A1 and the alkalis given by Carmichael e t al. (1974). Mueller & Saxena (1977). and Wilson (1989). the L C clast porphyry compares better with the acid derivates of continental tholeiitic flood basalts and other continental rhyolites than with oceanic and vol- canic arc rhyolites. Accordingly, the clast por- phyry is most probably derived from a continental source. Trace element data o n the L C porphyry clasts are given in Table 4. O h t a (1984, table 6) provided trace element data only for Sr, Ni, Co and C r for the Nordaustlandet quartz porphyries, classified as H-K and M-K C A rhyolites. Compared with his figures, Sr in the clast porphyry is similar t o the M-K C A rhyolite. but much higher than the H-K rhyolite. Ni and C r a r e much lower than i n both kinds of t h e Nordaustlandet rhyolites; Co is also lower. T h e content of PIOs is distinctly lower in the clast porphyry (Fig. 11A), and .the Ni/Cr proportions are also different (Fig. 11B). Trace Composition and provenance of the Lilljeborgfjellet Conglomerate 149 A j o t I I + I + a _ _ + - 0 0 -- - 0 I I 1 I I _-c- _--- 0 R h yo1 i le Dacire *-I_* I I ' J 65 70 7 5 SIO, w t % [ h y d r o u s ) B FeO' C a2 * / \ - - 4 - ,z / / / / / f I I I _- Or + + + Na,O + K,O Ab Fig. 8. A . Alkali-SiOz diagram (hydrous). B. AFM diagram. C. Normative Qz-Ab-Or diagram. Symbols: Open circles = Botniahalveya; solid squares = Sabinebukta; solid circles = Prim Oscars Land; crosses = Lilljeborgfjellet Conglomerate clast porphyry. Reference for dividing lines in Ohta (1984, fig. 9). 80 0 5 7 0 . B - a . 0 - + + + . / : ' + + 0 . ,' +.+ / f *+ 5 2 3 U 1 + _ , C 3 I L 1 ' + .+ a + + a ' ++ ++* + .'ct * * + - -+- . 0 c' 3 z 1 . 1 2 3 4 5 6 1 FeO' MgG Fig. Y. A Si02-FcO'Mg0 d i a g r a m . B. FcO-FcO \1gO diagram. C-. tlarker \ a r i a t i o n d i a s r a m s . Symbol5 as for Fig. 8 Refercncc f o r dividing lines i n O h t a (198-1. fig. 1 0 ) + + + I- $+ a + + m + ' + +*+* <> r * + element data of rhyolites elsewhere a r e scarce. even rarer for quartz porphyries. According to tables in the textbooks of Carmichael et al. (1971) and Wilson (1989). C r . Ni, Ba, R b and Zr in the LC clast porphyry compare well with those of the rhyolites in continental. tholeiitic flood basalt provinces. But Sr is somewhat lower, as a r e P 2 0 5 and TiO?. Discussion a n d conclusion The chemical and mineral compositions of the welded tuffs in the Frznkelryggen Sandstone For- mation differ considerably from that of the LC quartz porphyry clasts. According to the description by Murajov et al. (1983). the alter- ation of the tuffs also seems to be greater. I f they were to be the source rocks. it would also imply an inversion of the succession since the Siktefjellet Group underlies the R e d Bay G r o u p . As there is C "[ <> + 1 : + a + . + i r 5 t 70 SIO, 75 wi 46 (anhydrous) 80 Composition and provenance of the Lilljeborgfjellet Conglomerate 151 T d e 4 Tracc clrments of quartz porphjry clasts. Sarnplc no. Ba Co Cr Cu Nh Ni R h Sr Y Zn Zr 78-Gj.lllC 985 29 16 17 55 1.1 274 39 103 24 418 78-GJ.151A 1066 41 L 7 60 6.9 122 72 111 24 SO8 78-Gj. 157A 825 33 14 15.6 53 10.1 107 26 127 48 636 78-Gj.159A 695 37 20 7.1 49 5.7 125 21 111 30 628 78-GI. 159B 781 49 27 22.4 45 5 . 1 151 51 86 30 341 78-Gj.159E 1124 29 12 1.6 51 - 78-Gj.177D 1992 40 - 0.7 47 4.1 143 84 99 20 421 78-Gj.182 690 36 7 1.5 55 3.4 101 84 102 48 507 45 0.9 187 127 Y I 12 34Y 78-Gj.191D 776 26 60 - Averagc(x) 1331 40 19 6 53 4 153 h6 109 4 1 S1)h 7 7 8 4 ; . 151B 883 40 25 3.5 58 1.9 226 37 115 38 468 78-GJ.1579 1192 62 7 Y 58 8.6 125 100 132 26 676 78-Gj.157C 1124 36 18 5 . 5 57 2.6 99 93 128 116 684 78-Gj.lSYC 1274 57 13 4.3 48 4.8 154 66 104 67 548 183 61 I09 4 1 40.5 o n . I0 17 3 30 For sample locations see Table 3 . no indication of an inversion here, this alternative can be ruled out with some confidence. Bulk chemistry and petrography of the L C clast porphyry are in many respects similar to that of Nordaustlandet. T h e main petrographic dif- ference is the structure of the alkali phenocrysts and, at least in relation to the quartz porphyries of B and S , a more advanced recrystallization of the L C porphyry clasts. Both associations com- prise subvolcanic and pyroclastic units which have been subjected to metamorphism in greenschist facies. In most respects there is also good cor- relation between t h e chemical composition of the quartz porphyry clasts and that of the Nord- austlandet quartz porphyries, especially those of Botniahalvoya. But there a r e deviations par- ticularly in the amounts of Al, C r and Ni, which speaks against the possibility of the Nordaust- landet quartz porphyries being t h e source rocks of the L C quartz porphyry clasts. This alternative also raises other objections. Not only is the trans- port distance from Nordaustlandet to Haakon VII Land very long, but the direction is transverse to the general structural trend of the basement. which would probably create topographic barriers for river transport. Considering the large size of some of the clasts, it is especially difficult t o reconcile the idea of such a long transport Fig. 10. Alkalis vs. 100 K 2 0 / N a 2 0 + K 2 0 diagram. Symbols as for Fig. 8. Reference for dividing lines in Ohta (1984, fig. 12). 0 0 a a I 10 20 30 40 50 60 70 80 90 100 100 K,O Na.0 + K,O 152 T. Gjelsvik A A A A acid Andesite I Dacite Rhyolite b a S l C Andesite 0 . A t AA A A - 3 ,..O . o&oc 0 0 0 0 I . @ . . . . 0 ++ a + t * 2 +a - 50 ’ 0,3 ’ 0.2 . 0.1 k+.+4 , . . . . . . . . . . . . . . , 1 00 150 200 250 Cr (PPm) Fig. / I . A PIO,-S~O, diagram. Symbols: 9 = hasic dike rocks. Nordaustlandel; open circles = porphyrite and andcsitc, Nordaustlandet: solid circle\ = metadiabase. Nordaustlandel: solid triangles = quartz porphyries, Nordaustlandet; crosses = LC cla\t porphyry. Dividing lines from Ohta (1YX4. fig. 1 l a ) . B . Ni-Cr diagram. Symbols: open circles with horizontal lines = high K:O calc alkaline rocks open trimgle = medium K:O tholeiites: solid triangles = high K>O tholciitcs: crosses = LC clast porphyry. Refercncc for hrokcn curve. Ohta (19XJ. fig. 1 l c ) distance. On the other hand. the rounding of the clasts suggests that they are not derived from the adjacent basement rocks. The quartz porphyries in the Hornsund area are also unlikely to be the source of the LC clasts, since the chemistry as well as the petrology of the two types are v e r y different. T h e distance between them. 3UU k m , is also very long, but it might have been less prior to possible strike slip movement along the western Spitsbergen fold belt direction. Also from Nv Friesland ( N F on Fig. 1). the northeastern peninsula of Spitsbergen. rocks thought to be derived from acid volcanics have been reported (Gayer & Wallis 1966; Wallis 1969). They occur both in the basal part of the Planetfjella G r o u p and in the upper formations of the underlying Ilarkerbreen Group: they mostly comprise massive. feldspar-rich beds, now present as augen-gneisses. They are considered to be reworked. acid pyroclastics (Gayer & Wallis 1966). Some smaller beds consisting of crenulated schists with “megacrysts“ of feldspar ( u p to 5 mm in length). set in a fine-grained groundmass. are called meta-acid-tuffs. They have not been reported to contain relics of quartz phenocrysts. N o chemical data on these rocks has been published, and their origin is uncertain (Wallis 1969). Recent zircon U/Pb dating ( G e e 1991, unpubl. data from Peucat & Tebenkov 1990; O h t a in press) has shown that the Harkerbreen G r o u p belongs t o the Sveco-Karelian period, whereas the Planetfjella G r o u p may b e of Grenvillian age. The rocks have been subjected t o orogenic deformation and recrystallization in amphibolite facies, and the feldspar “megacrysts” occurring in both groups could b e a product of Caledonian or older events. Whatever their origin, they are not similar t o t h e LC quartz porphyry. Thus, none of the known outcrops of quartz porphyry in Svalbard satisfy the petrological and chemical criteria for the source rock. Most of the clasts in the LC can readily be matched with lithologies in the adjacent basement. O t h e r alternatives must therefore be considered, such as quartz porphyry bodies which have been com- pletely eroded. T h e adjacent area of northwestern Spitsbergen consists of amphibolite facies gneisses, migmatites and granite. A t originally higher stratigraphic levels, extrusive or sub- volcanic rhyolite might have occurred. As referred t o above, “welded tuff” of quartz por- Composition and provenance of the Lilljeborgfjellet Conglomerate 153 phyric composition is interbedded in the Fraen- kelryggen Formation of the Red Bay Group which was deposited in the early Gedinnian (Blieck et al. 1987) at about 400410Ma. However, the source rock in question would have to have been formed earlier. Acid igneous activity (including the Hornemantoppen granite) in northwestern Spitsbergen started somewhat earlier, a related vulcanite might have provided a source. However, the interrelationship of the deep-seated granites of north-western Spitsbergen and the welded quartz porphyric tuff in Frankelryggen is still an open question. Although the corre- spondence of most elements in the two rock types is good (Table l), it is less so for the alkalis and Mg. The correspondence is somewhat better for these critical elements when comparing the post- tectonic granites and porphyry of the LC clasts, but the A1203 contents of the former are more than 2% higher (Table 1). In 21 samples of syn- tectonic granites (> 68% S O z ) from the Smeer- enburgfjorden area, the average Al2O3 value is more than 3% higher (calculated from Ohta, table 5 , pp. 94-95 in Hjelle & Ohta 1974). In any case, because of the deep erosion, the tectonic setting, and the high metamorphic grade of the rocks to the north of Kongsfjorden, a possible source rock on the western side of the Devonian Graben could only be expected south of Kongsfjorden. Despite high mapping activity in recent years on the western coast of Spits- bergen, no quartz porphyry outcrop has been reported north of Bellsund. The same is the case for the anticlinal Hecla Hoek horst in the Raudfjord Devonian Graben. Other possible source rocks could be hidden under the sediments of the Devonian Graben, under large ice caps, or under the sea off the northern coast of Spitsbergen. So far, no quartz porphyry pebbles have been reported from the shelf, but very little sampling has been done. The northern shelf is perhaps not a probable source area, since it would imply a transport direction in Devonian time from north to south, opposite or transverse to what is believed by Orvin (1940, p. 15) and Friend & Moody-Stuart (1972, p. 17). The conclusions of their studies, however, are based on sedimentological investigations of the Red Bay/Wood Bay Groups, and the pale- ogeographical situation during deposition of the Siktefjellet Group might have been different. The seemingly increased degree of rounding south- wards of most clast lithologies in the latter indi- cates a source area to the north of the present location. This does not apply for the quartz por- phyry clasts, however. which are equally rounded everywhere. It may be concluded that the present infor- mation is insufficient to give a definite answer to the question of the provenance of the quartz porphyry clasts. Isotope dating of zircons from these clasts may provide a useful constraint on the potential alternatives. Acknowledgemenrs. - The chemlcal analyses have been carried out by S . Miyashita, Department of Geology, Faculty of Science. Niigata University. Japan. using x ray fluorescence methods. Y . Ohta and D . G . Gee have read the manuscript and provided valuable guidance. which does not relieve the author of full responsibility for the text. References Birkenmajer. K. & Narcbski, W. 1960: Precambrian amphibo- lite complex and granitization phenomena in Wcdel-Jarlsberg Land, Vestspitsbcrgen: Stitdin Geol. P o l . IV, 37-82, Blieck. A , . Goujet, D . & Janvier. P. H. 1987: The vertebratc stratigraphy of the Lower Devonian (Red Bay Group and Wood Bay Formation) of Spitsbergen: M o d f r n Geologv Vol. 11, 197-217. Carmichael. I . S . E.. Turner, F. J . & Verhoogcn. J. 1974: Igneous perrologv. McGraw-Hill. Flood, B.. Gee. D . G . , Hjelle A.. Siggerud, T . & Winsne,. T . S. 1969: Thc geology of Nordaustlandet, northern and central parts: Norsk Polarinst. Skr. 146. 141 pp. Friend, P. F. 1961: The Devonian Stratigraphy of north and central Vestspitsbergen: Proc. Y o r k s . Geol. Soc. 33. 77-1 16. Friend. P. F. & Moody-Stuart, M. 1972: Sedimentation of the Wood Bay Formation (Devonian) of Spitsbergen: Norsk Polarinst. Skr. 157. 77 pp. Gayer, R. A. & Wallis, R . H . 1966: The petrology of the Harkerbreen Group of the Lower Hecla Hoek of Ny Friesland and Olav V Land, Spitsbergen: Norsk Polarinst. S k r . 1.10. 32 PP . Gee, D. G . 1966: The structural geology of the Biskayerhoekcn Peninsula, North Spitsbergen. Ph.D. thesis, University of Cambridge (unpubl.). Gee, D. G . 1991: Caledonian Terranes on Svalbard. New evidence of basement in Ny Friesland. Abstract: Geonyrr drgang 18 (1191). p. 24. Gee, D . G. & Moody-Stuart, M . 1966: The base of the Old Red Sandstone on central north Haakon V11 Land, Vest- spitsbergen. N o r s k Polarinst. A r b o k 1964, 5 7 4 8 . Gorochov, J . M . . Kra'sil'Stikov, A . A . . Melnikov. N. N. & Varsavskaya. E . S. 1977: Rb/Sr age of quartz porphyries of the Kapp Hanstecn Series. Pp. 51-61 in P r o b l e m y geo- chronologii i geochtmii isotopou. NAUKA. Leningrad (in Russian). Harland, W. B. 1961: An outline structural history of Spits- bergen. Pp. 68-132 in Raach. G . 0. (ed.): G e o l o g y of the Arctic. Univ. of Toronto Press. Hjelle, A . & Ohta. Y . 1974: Contribution to the geology of north western Spitsbergen: N o r s k Polarinst. Skr. 158, 7-37. 154 T . Gjelsvik Holtedahl. 0. 1926. Soteh t i n the gcolog! of north-ucstern Spitshergcn: S k r . S i w l h . U K Idrocw. n r . 8. 2 8 pp. Lc M;iitrc. ed. lY8Y: .4 c.lrirsificcrrron of lgnuoiis R o d s ( i d Glo.\ror! of T r r t m Black\\cll Scientific Puhlications. Muellcr. R F 6: Sa\cn;i. S K . 1Y77 Geoclretnicol Perrulugr Springer-Verlag. MuraSov. L . G . . Pfelina. T . M. & Semevskij. D. V . 1983: N e w data on Evidence of Volcanism in Lower Devonlan and Upper Triassic Formations of Vestspitsbergen: Geology of Spilsbergen. Collection of Scientific p a p e r s , 96-101. Sevmor- geologija Leningrad (translated 1984). NordensttOld. (i I X V ? . Rcdogjor fdr dcn svcnaha cxpe- J i t i ~ e n t i 1 Spet5hcrgen lSY(l; Biliurig Kring. So Ver. . A k u f . HdI 17/.1/ I Y N : Gcochcnii\!r! of thc late Proterozoic Kapp tian\teen i g n c i w rocks o l N o r d a u d a n d e t . Svalhard. IYXJ: Polirr Keworch 3 t i . L , 6Y-02 Ohta. Y . in press: Recent understanding of the Svalbard base- Ohta. Y ment. in the light of new radiometric age determinations. Norsk G e o l o g i s k Tidsskrifr. Orvin. A . K . 1940: Outline of the geological history of Spits- bcrgen: N o r s k Polurinsr. Skr. 78. 57 pp. Peucat. J . J . . O h t a . Y . . Gee. D . G . di Bcrnard-Griftithb. J . I Y X Y . U-Ph. Sr and Nd cvidencc of Grcnvillian and latest Proterozoic tectonotherrnal activity in Spitdxrgen Calc- donidel. Arctic Ocean. Lirlio~ 2 2 . 175-285. Srnulikowski. W . 1Yh8: Some petrological and structural ohser- \ations in the Hecla Heel, succcsticn hetwccn Wer- enskioldhrccn and Torellbreen. Vest qitshergen: S r u d i c r Geologicrr Polotricci. Vol. X . Y I . 97-1.51 h ’ a I l i 5 . R. H . l Y 6 Y : T h e Planetfjclla Group of the Lower Hrcla Hock of N y Fricsland. Spitshergcn: Norvk Pol(irrfrst Arbok 1967. 7Y-lox. Wilson. M. 10x9: Igneous petrogenesis - a global tecronic npproacli. Unwin Hyman.