<<

134 Nature Vol. 269 8 September 1977

IO Nagy, 8 . Carbonaceous Meteoritr.'i (Elsevier, Amsterdam, 1975). during the ascent of . The lateral variation outlined 11 Woeller, F . & Ponn:imperuma, C . l carus 10,386 (1969). 12 Khare R. N. & Sagan C . Icarus 20, 311 (1973). above may be due to the lateral change in physicochemical 1J Scattergood, T., Lesser, P. & Owen, T. Icarus 24,465 (1975). 14 Herbig, G. H. Mem. Soc. Roy. Sci. Liege Ser. 5, 19, 13 (1970). condiitions of magma reservoirs across the volcanic zone. 15 Knacke, R. F. Na/Ure 217, 44 (1968). It 16 Zaikowski, A. & Knacke, R. F. Astrophy .,. & Space Sci. 37, 3 (1975). may also be due to a lateral variation in the chemical 11 Gillen, F . C., Kleinmann, D. E., Wright, E. L. & C_apps, R. W. Astrophys. l. Lett. composition of the odginal , for example, their 198, 1.65 (1975). . "Meinschein, W. G., Nagy, B. & Henessy, D. J. A1111. N. Y. Acad. Sc,. 108, 553 alkali contents---especi:al,ly the K,O contents' and possibly to (1963). the variation of H2O contents in the original magma which may increase westwards across the volcanic zone. I thank Profossor I. Kushiro and Dr R. H. Grapes for their comments and criticisms. Lateral variation of phenocryst assemblages in volcanic M. SAKUYAMA Geological Institute, rocks of the Japanese islands University of Tokyo, To understand the origin of calc-alkaline rocks and island Tokyo, Japan arc volcanism, knowledge of the lateral variation of the petrological characters of volcani;c rocks is es;.e.ntial. The Received 31 May; accepted 5 July 1977. lateral variation of chemical compositions of volcanic rocks 1 Kuno, H. Bull. /. 29, 195 (1966). 2 Kuno, H. Bull. Volcano!. 32, 141 (1968). has been investigated by several authors'-' and correlated 3 Hatherton, T. & Dickinson, W. R. l. geophys. Res. 74, 5301 (1969). 4 Sugimura, A. Bull. Volcano/. Soc. lap. 2nd Ser. 4, 77 (1959). to the depth of the Benioff zone. The east volcanic zone of 5 lsshiki, N., Matsui, K. & Ono, K. Geo/. Surv. lap. (1968). Japanese Islands, which contains about 140 discrete Quater,nary volcanoes, is ,the moot e,ctensive.ly studied volcanic zone in the world. l have examjned the petrological data of these volcanoes, particularly with respect to the Rare earth element mobility phenocryst assemblages in volcanic rocks and their changes during the history of respective volcanoes. The references and geochemical have been mosLly taken from 200 pa,pers compiled by Isshtkj characterisation of spilitic rocks et al.'. HELLMAN & Henderson 1 have demonstrated the enrichment of The volcanoes are classified i,nto three groups on ,the basis the rare earth elements (REE) in the lower spilitic portion of' the of the presc.nce or absence of hornblende and biotite Bhoiwada lave profile relative to the fresh tholeiiric flow top. The phenocrysts in ,essential mat,er:ials: (1) those without bio:ti,te remarkable feature of the REE distribution in the spilites was !he and hornblende phenocrysts in any of the or pyro­ absolute enrichment of all the clements and not only the light REE, clastics; 2) those with hornblende phenocrysts but without as occasionally seen in oceanic due LO suhmarine weather­ biotite phenocryst at least in one flow or pyroclastic ing2. Thus the enhanced spilite REE destribution resembles either rock; and (3) those with hiotite phenocrysts at least in one alkali or tholeiitic differentiation products rather than Jav-.t or pyroc-lastic rock. The distribution of the vok:anoes 1 tholeiitic hasalt. As stated , the high degree of REE mobility of these three grnups is shown in Fig. 1. suggests that these elements must be used with caution in Volcanic rocks contaiining biotite phenocryst<; were determining the initial magma type or volcanic environmenf of erupted only ~n .the western par,t of the volcanic zone, ancient spilitic rocks. whereas those whioh do not contain any hydrous rni11eral At face value the REE distribution in thespilites is vcrysimilarto phenocrysts were erupted in the eastern part of the volcanic 3 4 that exhibited by the more alkaline units of the Deccan basalts · . wne. Volcanoes in a belt situated between the above two This feature suggests the possibility that the Bhoiwada profile may be,l,t:5 erupted rocks containing only hornblende as hydrous not be a single flow unit at all and the fresh tholciitic 'top' may be a phenocryst. different unit to the more 'alkaline' spilite base. Although the Phenocrysts such as pyroxene, hornblende, and biotite section is described as gradational 5·6, there are primary textural are considered to be predpitated Ln a magma reservoi,r or differences (piilowed/massive; ophitic nature; vesicularity and so on) between tholciite and spilite, and the contact seems sharp (sec Fig, ) Distribution of Quaternary volcanoes in Hokkaido and photomicrographs and photograph in refs 5 and 6). If the sub­ northern Honshu. e, Volcanoes without bioti~e and hornblen_de phenocrysts in any of the lavas or pyroclast1cs; () those with aqueous (flow 'base') sub-aerial (flow 'top') origin postulated for hornblende phenocrysts but without biotite phenocryst at least the flow 5 is correct, then it lacks some of the features described for 7 8 in one lava flow or pyroclastics; open circles represent t)lose such lava flows · and does not adequately explain why spiliti­ with biotite phenocrysts at least in one Java or pyroclast1cs. sation should have stopped abrupted at the water-air interface. Also the considerable REE distribution gap (Fig. 2a or rer. l) between tholeiite and spilite might lend support lo the differences in the two zones, especially as there is no apparent gradational increase in normalised REE patterns with progressive spilitisation. This is particularly noticeable in the sudden jump of the La/Yb ratio 1 from tholeiite to spilite and also the chemical similarity between the spilite and the (separate) basal pillow lava unit. Irrespective of whether the profile represents a single unit or not, however, independent criteria are required to characterise the spilite and compare it with the fresh thokiite. A variety of geochemical (and mineralogical) parameters, which will give the same characterisation, are necessary to determine the nature of 6 9 ancient altered volcanics. Vallance • has advocated the use of relict pyroxenes to determine the magma type of spilitic rocks. In the case considered here, the pyroxenes in both tholeiite and spilite arc low-Ca augitcs of tholeiitic character and are similar in 6 composition . Apart from the REE, other relatively 'immobile elements', such as Ti, P, Zr, Y and Nb, have been used to determine 10 1 1 magma type • and again demonstrates the tholeiitic nature of

© 1977 Nature Publishing Group