Geochemical Journal, Vol. 36, pp. 133 to 153, 2002

Petrochemical study of the Jingpohu alkali basaltic rocks, northeastern China

ZHAOCHONG ZHANG,1* CHENGYOU FENG,2 ZHAONAI LI,1 SHUCAI LI,3 YING XIN,3 ZHAOMU LI3 and XIANZHENG WANG3

1Institute of Geology, Chinese Academy of Geological Sciences, Beijing 100037, China 2Institute of Mineral Deposits, Chinese Academy of Geological Sciences, Beijing 100037, China 3Institute of Geological Sciences of Heilongjiang Province, Ha-erbin 150036, China

(Received March 12, 2001; Accepted January 7, 2002)

Holocene volcanoes in the Jingpohu region are situated in the “Crater Forest” and “Frog Pool” areas along the northwest side of Jingpohu Lake, northeastern China. Dating of three charcoal samples from the first and second volcanic cycles shows that the ages of the first and second cycles are 3430~3490 and 2470 years respectively. The from the Jingpohu area consist of (BSN), alkali basalts (AOB) and (TP). Crystal fractionation models are consistent with the generation of AOB and TP from a basanitic parent. Minor fractionation of olivine, augite, magnetite and Cr-spinel is required to produce AOB compositions whereas the generation of TP requires extensive fractionation of kaersutite, phlogopite and anorthoclase with minor olivine, augite, magnetite and leucite. The presence of kaersutite, phlogopite and anorthoclase megacrysts and mantle xenoliths suggest a fractionation history occurring at high pressure in the mantle. Although all basaltic rocks contain many granitic xenoliths, their geochemical characteristics show that they have not undergone any contamination of upper en route to the sur- face, but some alkali basalts were suffered from the contamination of lower crust. Relatively unradiogenic isotope ratio (compared with Bulk Earth), steep chondrite-normalized REE patterns and strong incompat- ible element enriched patterns suggest that the are derived from a mixture of an incompatible element depleted anhydrous lherzolite asthenospheric mantle source and an enriched, -phologite- (apatite-)bearing lherzolite continental lithospheric mantle source. We propose that the basanites are the products of very low degree partial melts (<1%) of this source under high extension strength.

processes (e.g., Hofmann, 1988). On the other INTRODUCTION hand, there is increasing evidence that melts origi- The role of the lithospheric mantle in the nating from the asthenosphere interact with the petrogenesis of continental alkali basalts is diffi- lithospheric mantle en route to the surface (Chazot cult to evaluate, as its nature and composition are et al., 1996; Wulf-Pedersen et al., 1996; Class and known to be extremely variable, not easily distin- Goldstein, 1997). Current debate concerns whether guishable from the asthenospheric mantle. Conti- continental basalts are primarily derived from the nental alkali basalts have distinct geochemical sig- asthenospheric mantle, and contaminated by natures compared with mid-ocean ridge basalts lithospheric mantle or if they originate from vari- (MORB), but exhibit many similarities to oceanic ably “metasomatized” lithospheric mantle (e.g., island basalts (OIB). Lithospheric contamination Menzies and Hawkesworth, 1987; Beccaluva et of the asthenospheric mantle may occur through al., 1991; Fancis and Ludden, 1995; Wulf- lithospheric delamination (e.g., McKenzie and Pedersen et al., 1996; Comin-Chiaramenti et al., O’Nions, 1995) or through ancient subduction 1997; Simonetti et al., 1998; Marzoli et al., 2000).

*Corresponding author (e-mail: [email protected])

133 134 Z. Zhang et al.

gion. occurred frequently from Eocene to Holocene. The Holocene volcanic rocks are dis- tributed in two districts named “Crater Forest” and “Frog Pool” (Fig. 2). The major rock types include alkali olivine basalts (AOB), basanites (BSN) and tephrites (TP). AOB and BSN occur at “Crater Forest”, whereas TP occurs in “Frog Pool”. Based on interlayed sediments, three cycles can be iden- tified. All three cycles are found at “Crater For- est” and two cycles are found at “Frog Pool”. Although a lot of geochemical data for Terti- ary to basalts in the eastern sector of the Jingpohu region have been reported (e.g., Peng et al., 1986; Zhou and Zhu, 1992; Tu et al., 1992; Chen and Xu, 1992; Hsu and Chen, 1998), Holocene basalts discussed here have not been studied in detail. Here we present new 14C, major and trace element, and Sr, Nd and Pb isotopic data, and discuss the implications of those data in terms of petrogenesis.

CENOZOIC MAGMATISM IN JINGPOHU

Cenozoic basalts in Jingpohu are located at the west side of the Tangcheng-Lujiang (Tan-Lu) fault, which extends for ~2400 km in China and Fig. 1. Map of East China showing the main tectonic continues northeastern into the Russia (Fig. 1), and domains, distribution of Cenozoic volcanic rocks (modi- fied from Fan and Hooper (1991)), and location of Fig. left-laterally horizontally displaces several hun- 2 (square). Abbreviation: XM = Xing’an-Mongolian dred kilometers. The Jingpohu Cenozoic basalts fold belt; NC = North China (Sino-Korean) craton; cover an area of ~15000 km2 with a total volume QL = Qinling fold belt; YZ = Yangtze craton; SC = of >50 km3 and extending nearly 350 km north- South China fold belt. south as a part of Cenozoic belt of eastern China. The magmatism occurred in six episodes, at Eocene (44.9~42.1 Ma), Middle Miocene (16.49~8.65 Ma), Early Pliocene (4.09~3.36 Ma), An opportunity to determine the contribution Late Pliocene (2.6~2.4 Ma), Pleistocene (1.17~0.9 of lithospheric mantle to the petrogenesis of con- Ma) and Holocene (3490~2470 years). Cenozoic tinental alkali basalts is presented by the Jingpohu magmatic activity commenced in the Eocene with Holocene volcanoes, which are formed in eruption of tholeiites, forming two small Heilongjiang Province of northeastern China. The subhorizontally bedded sheets, orientated along a Jingpohu volcanic rocks constitute part of the NE-SW trend. A small scale of eruption along a Cenozoic belt of eastern China (Fig. 1), which is NE-SW trend occurred in Middle Miocene, com- an important component of the circum-Pacific posed of several low platform. The Early volcanic belt. The volcanic sequences were dis- Pliocene episode is the major magmatic event in tributed in northeast-southwest trending basins the region, with eruption of alkali olivine basalts, that parallel major tectonic structures in the re- some of which have minor iddingsite rims on The Jingpohu Holocene alkali basaltic rocks, northeastern China 135

Fig. 2. Simplified geological map of the Jingpohu region showing the exposure of Cenozoic volcanic rocks in the β region (modified from Heilongjiang Bureau of Geology and Mineral Resources, 1991). Q4 - Holocene basalt; β 2 β 1 β β Q3 - Pleistocene basalt; Q2 - Late Pliocene basalt; N2 - Early Pliocene basalt; N1 - Middle Miocene basalt; βE - Eocene basalt. Note: some small craters are not shown for clarity.

olivine, forming high lava platform in an area of and extends discontinuously 40 km) were formed ~1000 km2. Late Pliocene lavas accumulated within the lava flows. above valley, forming the third terrace. The Except Eocene basalts that belong to tholeeites, Pleistocene magmatic activity formed several those of all other episodes are sodic alkali basalts. small volcanic cones, including basaltic breccias However, compared with the alkali basalts of the intercalated with lavas. These cones also trend a other episodes, the Holocene alkali basalts are NE-SW direction. Holocene basalts are located at richer in alkali. Based on their geochemical and the west side of the Jingpohu volcanic region (Fig. Sr, Nd and Pb data, Liu et al. (1989) suggested 2). They are controlled by NE- and nearly W-E- that tholeiites formed by at a depth striking structures overlapped by nearly N-S-strik- of ~35Ð50 km, within the base of the plagioclase- ing structures, and distributed over large areas of lherzolite and the top of the spinel-lherzolite zone the Hercynian-Yanshanian (Mesozoic) granite ter- of the upper mantle whereas the alkali basalts ap- rain. Thirteen composite craters have been recog- pear to have formed at a depth of ~50Ð60 km en- nized. The craters consist dominantly of thin lay- tirely within the spinel-lherzolite zone of the up- ers of basaltic lavas alternating with volcanic per mantle. bombs, cakes, and scorias. The last stage of lava flows flowed for more than 60 km eastwards along ANALYTICAL METHODS a river valley and blocked the river, thus forming some barrier lakes, such as the well-known Analyses of major elements, REE and other Jingpohu Lake and Xiaobeihu Lake. Long lava trace elements were obtained from Institute of channels (about 1~3 m in width, 0.5~3 m in height Rock and Mineral Analyses, Chinese Academy of 136 Z. Zhang et al.

Table 1. 14C isotopic measurement for the Holocene volcanic rocks from the Jingpohu area

Sample No. Location Sample Age (B.P.) Correction (A.D.)

P33C1451 No. I crater charcoal 4630 ± 60 3430 ± 80 DG-8 No. V crater charcoal 3950 ± 70 2470 ± 120 DG-8 No. V crater charcoal 3970 ± 70 2470 ± 110 TC31 Frog Pool charcoal 4660 ± 80 3490 ± 140

Note: the used half life of 14C is 5568 years; initial age is 1950 years; liquid scintillation meter is Quantulus-1220; all the samples were determined by Yin Jinhui in Earthquake Bureau of China.

Geological Sciences (CAGS). Analyses of Sr, Nd Earthquake Bureau of China, Beijing. The method as well as Pb isotopic compositions were carried has been detailed by Stuiver and Reimer (1993), out at the Institute of Geology, Chinese Academy and results are given in Table 1. of Sciences. The analyses of major elements were performed by XRF on a Philips PW2400 auto- PETROGRAPHY mated spectrometer. The REE plus Rb, Sr, Y, Sc, Cr, Ni, Ba, Th, U, Zr, Hf, Nb, Ta were analyzed In this section, all the basaltic samples are by inductively coupled plasma-mass spectrometry fresh, do not contain any secondary minerals, and (ICP-MS, Perkin-Elmer-Sciex Elan Model 500) are mostly strongly porphyritic. Glass is present under conditions outlined by Dulski (1994). Un- in the samples of the chilled border of many lava certainties for Nd, Sm, and Sr concentrations were flows, and these samples usually contain 1Ð2% 0.5%. Analysis of radiogenic isotopes has been olivine and augite . The lava flows discussed by Harmer et al. (1986), and only a sum- have two occurrences, i.e., thin lava flow and thick mary is presented here. 87Sr/86Sr and 143Nd/144Nd flow. The thin lava flows generally show a high ratios were determined on a VG 354 mass vesicularity of 20 to 40 vol.%, generally 8Ð25 cm spectrometer, and isotopic ratios were normalized thick, and contain about 1Ð3% phenocrysts. The to 145Nd/144Nd = 0.7219 and 86Sr/88Sr = 0.1194. thick lava flows are usually gray, compact in out- Repeated analyses of standards yielded averages crop area, generally 0.8Ð2 m thick. These lavas of 0.710245 ± 0.000018 (2σ, n = 6) for Sr stand- are strongly porphyritic. Only compact lavas are ard NIST SRM987, and 0.511870 ± 0.000018 (2σ, mainly used for geochemical studies in this pa- n = 6) for the LaJolla Nd standard. Total chemical per. blanks were <500 pg for Sr and <100 pg for Nd. Mantle xenoliths can be usually observed in Pb-isotope ratios were measured on a VG MM30 the lavas and volcanic bombs. AOB and BSN can mass spectrometer using an optical pyrometer to be petrographically distinguished from TP, in monitor filament temperature to ensure constant which the latter often contains kaersutite, fractionation effects. Mass fractionation correc- phlogopite and anorthoclase megacrysts, while tions of 0.09 (±0.03, 2σ) %a.m.u. and 0.24%a.m.u. AOB and BSN lack any megacrysts. Kaersutite were applied to Pb isotopic ratios (based on re- and phlogopite megacrysts usually occur at the peated measurements of NIST SRM 981 Pb stand- lower part of volcanic sequence (in lavas and ard) using Faraday and Daly analogue detectors, scorias) whereas anorthoclase megacrysts occur respectively. Total procedural blanks are ≤100 pg. in the upper part, and enclosed in lavas. Kaersutite Pb isotope analyses followed the procedures de- megacrysts are common as black crystals, gener- scribed by Gulson et al. (1984). ally 0.5Ð1 cm, up to 7 cm long, some of which are 14C Measurements on charcoal were carried out prismatic with well developed crystal faces and in the 14C Laboratory of the Institute of Geology, perfect {110} cleavage. Phlogopite megacrysts are The Jingpohu Holocene alkali basaltic rocks, northeastern China 137

e as for this table.

ces of the same Sample numbers in other tables ar

ocks from Jingpohu ocks from

BSN-; AOB-alkali olivine basalt; TP-. FeOt = total FeO content. Data sour BSN-basanite;

Table 2. Major element analyses and CIPW norms of Cenozoic alkali basaltic r Table 138 Z. Zhang et al. usually euhedral in black brown, 0.8Ð2 cm, up to minor leucite. The groundmass consists of 5 cm in diameter. Anorthoclase megacrysts com- plagioclase, clinopyroxene, opaques and olivine, monly occur only as transparent single crystals, all smaller than 0.05 mm in diameter, with up to 0.5 to 1 cm, up to 2 cm long. 20 vol.% of brown interstitial glass. AOB and BSN have porphyritic textures and are always hypocrystalline. The most common 14C DATING OF THE HOLOCENE BASALTS phenocrysts are olivine and augite, with subordi- nate Ti-magnetite and plagioclase. The Three charcoal samples were used to constrain phenocrysts are up to 10 mm across and augite is the ages of Holocene basalts in the Jingpohu re- usually more common than over olivine. Only in gion. Sample P33C1451 was collected from scoria the most primitive lavas do olivine crystals make at the north side of No. I crater (Fig. 2). Sample up ~60% of the phenocrysts. Augite is euhedral DG-8 was sampled from the scoria of No. V cra- to subhedral and pale pink to purple in plane-po- ter, which belongs to the second cycle of larized light with a darker rim. Olivine phenocrysts volcanism based on field observation. Sample are generally euhedral to subhedral and 0.2 to 0.4 Tc31 was collected from scoria at the southeast mm in diameter. The groundmass consists of side of “Frog Pool” (crater No. XII), which corre- plagioclase laths and clinopyroxene grains less sponds to the first eruptive cycle of the “Frog than 0.1 mm, with intergranular opaques or inter- Pool” area. stitial glass and minor nepheline. Table 1 shows that charcoals from craters I, V, The tephrites are also porphyritic. The lavas and XII were buried 3430, 2470 and 3490 years are vesicular (5Ð40 vol.%) and contain less than 5 ago respectively, which shows that the age of vol.% phenocrysts in an interstitial groundmass. scoria from No. I crater is nearly the same as that phases are idiomorphic olivine and from crater XII (“Frog Pool”). It is also inferred

Fig. 3. Diagram of Mg# vs. selected major oxides and trace elements for Holocene alkali basaltic rocks from Jingpohu. Also shown: crystal fractionation trends and fractionating phases. The Jingpohu Holocene alkali basaltic rocks, northeastern China 139 that volcanic eruption in both regions took place (Fig. 4). at about the same time. The REE are strongly enriched relative to chondrite (Fig. 5), as might be expected from the alkalinity of the lavas. AOB and BSN are in the GEOCHEMISTRY range of 130 times chondrite for La, whereas TP Major elements reach an enrichment of ~200 times chondrite. The Representative major element analyses of the light REE (LREE) are strongly fractionated from Holocene Jingpohu basaltic rocks are presented the heavy REE (HREE), the latter having contents in Table 2. Compared with other episodes of 2 to 5 times chondrite in the case of Yb and Lu. basalts, Holocene basalts have higher total alkali (La/Yb)N ranges from 20 to 29 for the BSN and contents. For Holocene basalts, AOB and BSN are 45 to 50 for TP. Compared with those of other characterized by higher MgO, CaO contents, and episodes, Holocene basaltic rocks are more lower Al2O3, Na2O, and K2O contents in compari- strongly enriched in LREE (Table 3). All the son with tephrites. analyzed samples show very slight positive Eu # As the range of SiO2 contents is restricted, Mg anomalies (Table 3 and Fig. 5). value is used as an index of differentiation. In Fig. The strong incompatible element enrichment 3, TP, AOB and BSN fall in two fields according is also visible in the primitive-mantle normalized to Mg# values of 0.60 to 0.68 and 0.43 to 0.50. plots (Fig. 5, primitive mantle values from # For high Mg ranges, Na2O, Cr and Ni decrease Hofmann (1988)). AOB, BSN and TP display a # with increasing Mg values, while Al2O3 and CaO closely similar pattern. All lavas show positive contents increase. In contrast, at lower Mg# val- anomalies for Th, K, P and negative anomalies for ues, most major elements show at similar concen- Ti compared with Ba, U, Ta, Nd, Hf, Sm and Tb. trations. In addition, there are no Sr anomalies. However, only in terms of the highly incompatible element Trace elements Ba are AOB and BNS patterns distinguishable Compared with TP, AOB and BSN have higher from TP, in that AOB and BNS have positive contents of the compatible elements Ni and Cr, anomalies, whereas TP shows negative anomalies and moderately compatible elements Y, V, Co and (Fig. 6). In addition, Fig. 6 also shows a similar Sc, but lower contents of incompatible elements pattern with OIB, but with relatively high concen- trations of Rb, Ba, Th, and K.

Fig. 5. Chondrite-normalized REE patterns for Holocene basaltic rocks from Jingpohu region (only Fig. 4. Diagram of Mg# vs. Y, V, Co and Sc for Holocene some representative samples are chosen because of alkali basalts from Jingpohu. Symbols are as for Fig. 3. similar REE data). 140 Z. Zhang et al.

ea

ocks from the Jingpohu ar ocks from

Table 3. REE analyses of Cenozoic basaltic r Table The Jingpohu Holocene alkali basaltic rocks, northeastern China 141

ea

ocks from the Jingpohu ar ocks from

Table 4. Trace element analyses of Cenozoic basaltic r 4. Trace Table 142 Z. Zhang et al.

Fig. 6. Incompatible element patterns normalized to primitive mantle (Sun and McDonough, 1989) for Holocene basalts from Jingpohu. Only some representative samples are chosen because of similar data; symbols are as for Fig. 5. Black squares represent average OIB composition (Wilson and Downes, 1991).

Fig. 7. Diagram of 143Nd/144Nd vs. 87Sr/86Sr (a) and 87Sr/86Sr vs. 206Pb/204Pb (b) of Holocene alkali basaltic rocks from Jingpohu. Other symbols are as for Fig. 3. EMI and DMM are mantle end-member compositions from Zindler and Hart (1986), and MORB and OIB data from Wilson (1989). Other data sources are from Liu et al. (1989).

Sr, Nd, Pb isotopic data of emplacement, so age corrections are negligi- Representative Sr, Nd and Pb isotopic analy- ble. Thus, age corrections have not been made ses of the Holocene basaltic rocks in the Jingpohu here. Although the major and trace element com- region are presented in Table 5. Given the young position of AOB and BNS are different from TP age of all rocks, the measured isotopic ratios (Tables 2, 3 and 4), they are isotopically relatively closely approximate to the initial ratios at the time homogeneous (Table 5), suggesting that they origi- The Jingpohu Holocene alkali basaltic rocks, northeastern China 143

Table 5. Sr, Nd and Pb isotopic data of Cenozoic basaltic rocks from the Jingpohu area

isotopic ratios and lower Sr isotopic ratios. In con- trast, they have lower Pb and Nd isotopic ratios and similar Sr isotopic ratios in comparison with the Middle Miocene and Pleistocene alkali basalts.

DISCUSSION Crustal contamination It is necessary to evaluate the extent of crustal contamination before using geochemical signa- Fig. 8. 208Pb/204Pb vs. 206Pb/204Pb for Holocene alkali tures of the Jingpohu Holocene alkali basaltic basaltic rocks from Jingpohu. Data sources: Dupal OIB rocks to constrain their potential mantle sources. from Hamelin and Allegre (1985) with additional data Crustal contamination may be important in conti- from Hart (1984) and Hofmann (1997). Symbols are as for Fig. 7, and other data from Liu et al. (1989). nental petrogenesis (e.g., Haase and Devey, 1994; Wilson et al., 1995; Wenrich, 1995), despite the high ascent velocity of the parental magmas. The presence of granitic xenoliths in the nated from the same source region. The 87Sr/86Sr basaltic rocks suggests that some magmas may be ratios vary from 0.703968 to 0.704537 with an contaminated by the granite derived from base- average value of 0.704313, which is higher than ment. However, the all xenoliths have not been those of N-MORB but similar to those of some resorpted, implying that upper-crustal contamina- continental basalts. 143Nd/144Nd values range from tion was probably negligible. 0.512656 to 0.512786 with an average value of Thompson et al. (1984) have suggested that ε 0.512696, Nd +0.4 to +2.9, which is lower than the La/Nb ratio might be a useful index of crustal that of N-MORB (Viereck et al., 1989). 206Pb/ contamination in magmas, and Taylor and 204Pb, 207Pb/204Pb, and 208Pb/204Pb ratios (206Pb/ McLennan (1995) regard the ratio Nb/U as an in- 204Pb = 17.086~17.503, 207Pb/204Pb = dex of crustal contamination. All the analyzed 15.269~15.346, 208Pb/204Pb = 37.100~37.436, av- samples have La/Nb < 1 and high Nb/U ratio of erage values 17.369, 15.306, 37.273 respectively) 15 to 34 (Table 4), suggesting that they have not are lower than those of N-MORB and island arc undergone contamination by rocks of the conti- basalts. In Figs. 7 and 8, we note that the Sr, Nd nental crust. However, the observed isotopic and Pb isotopic compositions are plotted between geochemical variation (Table 5) on BSN and TP DMM and EM1. could reflect crustal contamination or mantle Compared with the Eocene tholeeites, source heterogeneity. Holocene basaltic rocks have higher Pb and Nd Some samples with high initial 87Sr/86Sr 144 Z. Zhang et al.

original . Thus, negative correction be- tween 87Sr/86Sr and 206Pb/204Pb and high U, Th concentrations fit a model dependent upon lower- crust contamination rather than upper-crust con- tamination. Estimation of the components involved (i.e., uncontaminated primary magma and contaminant) is probably the most difficult problem in modeling contamination processes. We have selected a primitive sample with the lowest 87Sr/86Sr ratio (DX1-26) to represent the uncontaminated magma, and the one with the highest 87Sr/86Sr ratio (JP-3) to represent the contaminated magma. As the most likely crustal contaminants were lower-crustal granulites, we have modeled crustal contamina- tion processes using compositional and isotopic Fig. 9. Diagram of La/Nb vs. 87Sr/86Sr for Holocene data for granulite-facies rocks at locations geo- alkali basaltic rocks from Jingpohu. Data sources: ex- graphically close to Jingpohu (Shen et al., 1991). cept for Sr isotopic ratio of continental crust from According to the formula of two end-member mix- Zartman (1984), other data from Weaver (1991). ing proposed by Faure (1986):

t t t I Sr,m = [I Sr,AXAfA + I Sr,B(1 Ð fA)] 143 144 (0.70446Ð0.70454) and low initial Nd/ Nd /[XAfA + XB(1 Ð fA)]. (1) (0.51266Ð0.51267) may be indicative of crustal 87 86 t 87 86 contamination. Moreover, this increase in Sr/ Sr Whereas I Sr,A and XA are Sr/ Sr ratio and is associated with a decrease in 206Pb/204Pb (Fig. Sr content of lower crust at t time respectively; t 87 86 9). Davies and Macdonald (1987) has related it to I Sr,B and XB are Sr/ Sr ratio and Sr content of t 87 crustal contamination. If granitic contamination uncontaminated magma; I Sr,m is the mixing Sr/ 86 is the cause of the isotopic variation, their La/Nb Sr ratio; fA = A/(A + B), representing the per- ratios and SiO2 contents will be elevated as these centage of crustal materials. We can convert Eq. granitic rocks are characterized by high SiO2 (1) to have the following expression: (>70%) and La/Nb ratio (>1) (Wu et al., 1999). t t t t Thus, the negative correlation between La/Nb ra- fA = (I Sr,A Ð I Sr,m)XB/[(I Sr,A Ð I Sr,m)XA 87 86 t t tios and Sr/ Sr ratios (Fig. 9) and low SiO2 con- + (I Sr,m Ð I Sr,B)XB] = 1/(1 + P). (2) tents (<50%) could not have resulted from the contamination of granitic materials of upper crust. Whereas Presumably, in a simple mixing situation, if t t t t basalt with typical mantle composition becomes P = [(I Sr,A Ð I Sr,m)/(I Sr,m Ð I Sr,B)](XA/XB). (3) contaminated by lower crust, the basalt would acquire a less radiogenic Pb-isotopic character. If, Shen et al. (1991) has obtained the initial 87Sr/ on the other hand, contamination takes place in 86Sr ratio of the granulites in the adjacent area, t the upper crust, it normally would be expected to 0.70335 and their age, 2790 Ma. The I Sr,A value acquire a more radiogenic lead isotopic character after 2790 Ma can be calculated by the following (Zartman, 1984). In both cases the U, Th concen- expression given by Faure and Powell (1972): trations and 87Sr/86Sr ratios in the contaminated 87 86 ≈ 87 86 87 86 λ basalts should be appreciably higher than in the ( Sr/ Sr)t ( Sr/ Sr)0 + ( Rb/ Sr) t. (4) The Jingpohu Holocene alkali basaltic rocks, northeastern China 145

t From the Eq. (4), we can obtain the I Sr,A value, tle and country rocks, and fractionating phases. 0.71053, and its Sr content is 490 ppm (Shen et Although mantle xenoliths contain amphibole al., 1991). Thus, we can calculate the fA value from (pargasite) and phlogopite, their mineral compo- Eq. (2): 16%, i.e., the sample JP-3 resulted from sitions distinguish from those of the megacrysts the contamination of about 16% lower-crustal (Zhang et al., 2001). In addition, the compositions materials from the uncontaminated magma (sam- of amphibole and phlogopite hosted in the regional ple DX1-26). strata are also distinguishable from those of the megacrysts (Heilongjiang Bureau of Geology and Crystal fractionation Mineral Resources, 1991). Thus, the kaersutite and Three factors controlling basaltic rock com- phlogopite megacrysts could not result from the positions are: 1) heterogeneity of the source re- xenocrysts of either mantle or country rocks. Since gion; 2) variable degrees of partial melting of the neither mantle xenoliths nor country rocks con- same source region; 3) fractional crystallization tain any anorthoclase, anorthoclase is not either and contamination. xenocrysts of mantle or of country. Their similarity in isotopic composition of the Sheng et al. (1983) reported the Sr isotopic Jingpohu Holocene basaltic rocks suggest that they compositions of kaersutite and phlogopite, and originated from the same source region. The low showed the Sr isotopic equilibrium between these Cr and Ni contents of TP indicate a relatively megacrysts and their host basalts, suggesting a evolved magma, consistent with their lower MgO parent-daughter relationship, i.e., cognate origin. and Mg#. Low Mg# usually means that magmas The experiments on alkali basalt compositions have undergone fractional crystallization en route show that anorthoclase is a subsolidus mineral at to the surface (Cox, 1980; Prestvik and Goles, low pressures (Chapman, 1976). Therefore, 1985). Such fractionation will tend to increase the kaersutite, phlogopite and anorthoclase incompatible element concentrations in the basal- megacrysts crystallized from their host basalts. tic magmas relative to those of the more MgO- Variations of Mg# vs. major oxides and some rich primary magmas. Therefore, the petrography, trace elements in Fig. 3 can be best explained by major element and trace element variations as well a two-stage crystal fractionation model, i.e., from as their isotopic compositions suggest that differ- high Mg# (0.60 to 0.68) to low Mg# (0.43 to 0.50), entiation of the BSN, AOB and TP is compatible and from Mg# of 0.68 to 0.60. From high Mg# to # with fractional crystallization. low Mg , increase of Na2O can be attributed to As basanites are the only relatively primitive the fractionation of leucite whereas decrease of rocks (Mg# values = 0.68) found on the “Crater CaO can be attributed to the fractionation of

Forest” region, we chose to investigate the possi- clinopyroxene plus amphibole. Increase of Al2O3 bility that both TP and AOB could be derived from and decrease of Ni are consistent with the the same parent magma by different crystal fractionation of olivine, and decrease of MgO and fractionation processes. Whereas TP follow a more Cr can be attributed to the fraction of olivine, Si-undersaturated trend, AOB are more relatively clinopyroxene as well as Cr-spinel. Similarity, in Si-saturated than any possible parental basanites high Mg# range, i.e., 0.60 to 0.68, increase of

(Table 2). Al2O3 and CaO and decrease of Ni can be best It is necessary to evaluate possible explained by the fractionation or accumulation of fractionating phases before modeling fractional olivine, and decrease of MgO and Cr are consist- crystallization processes. At first, we discuss if ent with the fractionation of olivine and the megacrysts and phenocrysts are cognate. It is clinopyroxene. The detail explanation for the vari- no doubt that phenocrysts are fractionating phases ation is shown in Fig. 3. from their parent magma. There are three possi- Since there are different mineral assemblages ble origins of megacrysts, i.e., xenocrysts of man- in TP and AOB, it is not possible to model the 146 Z. Zhang et al.

Table 6. Least squares mixing for crystal accumulation or fractionation models

Mineral% - percentages of minerals required to separated or accumulate; Observed value-analyzed compositions of parental magma; Calculated value-theoretically calculated composition of parental magma in the light of separated phases and daugh- ter magma; ∑r2 = sum of the squares of the residuals. The compositions of phenocrysts and megacrysts were used in calcula- tions from Zhang et al. (2001). Ol = olivine; Aug = augite; Cr-Sp = Cr-spinel; Kaer = kaersutite; Phl = phlogopite; Anor = anorthoclase; Mt = magnetite; Leu = leucite.

whole range of MgO contents in terms of an in- 6. The appearance of kaersutite and phlogopite as variant fractionation assemblage. We have there- two fractionating phases probably reflects a higher fore calculated a two-stage model which, based H2O content in the magma. This may also explain on the petrography of the samples, takes olivine, the paucity of plagioclase observed in the augite, kaersutite, phlogopite, anorthoclase, mag- tephrites, as well as the lower proportion of feld- netite and leucite as possible fractionating phases. spar in the model calculations, as Yoder and Tilley

As no negative Eu or Sr anomalies are seen in ei- (1962) have shown that H2O and high pressure ther REE plots or mantle-normalized diagrams, suppressed plagioclase crystallization. Kaersutite plagioclase should not be considered a fractionation leads to (1) lower CaO/Na2O ratios fractionating phase. Fractionation modeling was in the tephritic magmas, so that any plagioclase performed using the GPP programs of Geist et al. which does crystallize has lower anorthite con- (1985). Despite the strong zonation of augite, the tents (Table 2), and (2) silica-undersaturation in calculations yield good results (Table 6). the resulting magmas (Wass, 1979). In the first step of the crystal fractionation model, olivine, augite, magnetite and Cr-spinel are Constraints on pressure of fractionation extracted from a relatively primitive basanite Using the normative mineral plot of Sack et (DX4-8 with 9.25% MgO) to produce AOB al. (1987) for alkali rocks, some constraints on P32Gs7 with 7.66% MgO (Table 6). the pressure during crystal fractionation can be The second fractionation step consists of gen- obtained. Absolute pressures cannot be defined, erating a tephrite (HM-11) from sample DX4-8. however, as P has evidently played an impor- HO2 Here the effects of kaersutite, phlogopite and tant role at Jingpohu, as witnessed by the anorthoclase fractionation become important. The kaersutite and phlogopite fractionation history results of the calculations are presented in Table discussed above. The effects of P were not HO2 The Jingpohu Holocene alkali basaltic rocks, northeastern China 147

reported as a subsolidus mineral at low-pressures in experiments on alkali basalt compositions (Chapman, 1976). The lack of correlation of Ni or Cr with Mg# (Tables 2, 4 and Fig. 3) does not preclude frac- tional crystallization. It may be that fractional crystallization continued until the density of the magma was such that it rose to the surface. This could be happened several times, producing erupted tephrites with more or less the same compositional characteristics.

Lithospheric and asthenospheric mantle sources Although Sr, Nd and Pb variations with the analyzed samples can be observed, they have some Fig. 10. Normative Di-Ol-Ne diagram showing the similarities, which indicate that they originated recalculated normative compositions of AOB, BSN and from the same source region. In addition, high- TP from the Jingpohu. The traces of the 1 bar (FMQ) pressure experiments on dry lherzolites (e.g., olivine + clinopyroxene + plagioclase cotectic and the Adam, 1990; Hirose and Kushiro, 1993; Green and 8Ð30 kbar olivine + orthopyroxene + clinopyroxene Fallon, 1998) suggest that the SiO concentration cotectic are from Sack et al. (1987). 2 of mantle partial melts are pressure and tempera- ture (i.e., melting degree) dependent, whereas to- tal FeO concentrations are mainly pressure de- considered by Sack et al. (1987) in their experi- pendant (i.e., increasing with pressure). Therefore, ments. Figure 10 shows that all rocks plot between we suggest that BSN and TP may have been gen- 8Ð10 kbar and 1-bar cotectic, suggesting that they erated at similar pressure, as their total FeO con- fractionated at relatively great depth. centrations are similar (Table 2). The previous experiments show that amphibole As the stated above, their primary mantle nor- and phlogopite megacrysts are generated at high malized patterns display similar characteristics pressure (Irving and Frey, 1984; Forbers, 1974; with OIB (Fig. 6), and their Sr, Nd, Pb isotope Chapman, 1976; Green and Ringwood, 1976; data are also plotted within the field of OIB (Fig. Edgar et al., 1976; Barton and Hamilton, 1982), 7), which imply that they probably originate from and Ti-rich amphibole has been shown to be a near the asthenosphere or (e.g., Fitton liquidus phase in experiments on alkali basalt and Dunlop, 1985). Two evidences suggest that compositions, being stable up to ~31 kb, 1100°C the Jingpohu Holocene alkali basalts cannot be (Irving and Frey, 1984). Based on the considered the expression of mantle plume, as for geobarometer of amphibole proposed by Hollister example, the Hawaiian islands: (1) Eocene to et al. (1987), the depth of generation for kaersutite Quaternary volcanism in the Jingpohu region do of the “Frog Pool” is estimated to be 12.53~12.56 not display a transport trace, i.e., younging trend kbar, corresponding to depth of 41 km (Zhang et along a direction (Fig. 2). (2) Magmatism related al., 1999). The geophysical data show a continen- to mantle plume is characterized by the great vol- tal crustal thickness of 38 km (Zhao et al., 1996), umes and high eruption rates, but the Jingpohu thus implying that crystallization started in the is absence of the scenario. Thus, the lithospheric mantle. However, anorthoclase has Holocene alkali basaltic rocks from Jingpohu with not been shown to be a stable phase in high-pres- OIB geochemical characteristics probably origi- sure experiments in alkali basalts, but has been nate from asthenosphere. 148 Z. Zhang et al.

Ormerod et al. (1991) regards the ratio Zr/Ba be interpreted as being caused by a mixture of as an index of whether primary magma is derived EM1 and DMM endmembers. In general, DMM from asthenospheric mantle or lithospheric man- is considered to be related to asthenosphere (like tle. All analyzed samples have Zr/Ba ratio >0.2, MORB). Although the genesis of EM1 is still de- suggesting that they resulted from asthenospheric bated, most researchers suggest that EM1 is re- mantle. This is in agreement with a theoretical lated to metasomatized continental mantle treatment of the magmatic consequences of lithosphere by small melt fractions very rich in lithospheric extension by McKenzie and Bickle potassium and volatiles released from (1988), i.e., the generation of alkali basalts in con- asthenosphere (Menzies, 1990; McKenzie, 1989; tinental rift zones must be confined largely to the Hawkesworth et al., 1990) or by fluids rich in convecting asthenosphere. Therefore, a dry, de- LILE released from dehydration due to the sub- pleted, lherzolite asthenospheric mantle source duction of oceanic plate (Hawkesworth et al., lying below the lithosphere must be considered 1990; Tatsumoto et al., 1992; Liu et al., 1994). as a potential source component for the Jingpohu As the Jingpohu region is too far away from sub- Holocene alkali basalts. However, an important duction zone (Peng et al., 1986), EM1 component implication of McKenzie and Bickle’s calculations may be related to metasomatized continental man- is that unless the thickness of the mechanical tle lithosphere by small melt fractions very rich boundary layer (MBL) is less than 65 to 80 km, in potassium and volatiles released from partial melting of asthenosphere is unlikely. The asthenosphere. thickness of the lithosphere in this area is about In Fig. 4, BSN show positive inflections for 100 km. This raises the question: how did partial Ba, K and P, and a negative inflection for Y, which melting of asthenosphere beneath the Jingpohu can mostly easily be explained by melting of area trigger? It is noted that their model assumes phlogopite, amphibole and apatite, with residual that the mantle sources involved are “dry” garnet in the source. Amphibole in peridotite peridotite, not “wet” peridotite. Considering the xenoliths from “Frog Pool” is characterized by conditions of “wet” lithospheric mantle and “dry” positive Ba and K anomalies, and by relatively asthenospheric mantle, Gallagher and enriched LREE concentrations (Lee et al., 1996; Hawkesworth (1992) revised the model proposed Marzoli et al., 2000). For positive K anomalies, by McKenzie and Bickle (1988). The revised there is evidence for amphibole control, but in model shows that asthenosphere can generate a most cases the potassic phases appears to be large amount of melt with the lithosphere of the phlogopite, interpreted to be the result of small thickness of 100 km at the stretching factor >1.2 amounts of phlogopite in their source regions to 1.3. This suggests that the Jingpohu Holocene (Hawkesworth et al., 1990), which also present alkali basalts generated under high extension in “Frog Pool”. Considering the low P-T stability strength. of both amphibole (<30 kbar and <1050°C; see On the other hand, as noted above, Holocene Mengel and Green, 1989; Foley, 1991), phlogopite primitive alkali basalts are relatively more en- and of apatite (Watson, 1980; Baker and Willey, riched in large ion lithophile elements (LILE) than 1992), and considering the lithospheric thickness OIB, which suggests that a fertile continental of the lithosphere in this area (100 km; Zhao et lithospheric mantle should be another potential al., 1996), the amphibole- and phlogopite-related major source component (Perry et al., 1987; chemical signature of the magmas suggests a Hawkesworth et al., 1990; Fitton et al., 1991; lithospheric mantle source component in the Wilson and Downes, 1991; Rogers et al., 1995). petrogenesis of basanites. This is also supported by their Sr-Nd-Pb isotopic The degree of melting over which basanite is compositions. All basaltic rocks are plotted be- produced at 20Ð35 kbar ranges from <1% to 6% tween EM1 and DMM, suggesting that they can (Green, 1973; Ringwood, 1975; Edgar, 1987). The Jingpohu Holocene alkali basaltic rocks, northeastern China 149

Mysen and Kushiro (1977) generated primary magma compositions and can be used for undersaturated magmas at 1Ð2% partial melting REE modeling. In this modeling, incompatible at 20 kbar and 1450Ð1475°C from garnet element depleted garnet lherzolite from McKenzie peridotite. These melts had approximately and O’Nions (1991) is chosen to represent for basanitic chemistry although TiO2, Na2O, and K2O asthenopheric component, and fertile amphibole- were too low. Unfortunately, experiments at these phlogopite-garnet lherzolite for lithospheric com- low degrees of melting have so far suffered from ponent. The least evolved samples fall between serious technical problems with quenching and the equilibrium batch melting curves for a depleted equilibrium. garnet lherzolite and an amphibole-phlogopite- Some analyzed basanite samples with high Mg# garnet lherzolite (Fig. 11a). Using a mixture of values have compositions that approach those of 90% depleted garnet lherzolite and 10% primary melts. Incompatible trace element ratios amphibole-phlogopite-garnet lherzolite, about and Sr, Nd and Pb compositions of these samples 0.5% melting will produce the relatively primi- can provide some constraints on their mantle tive basanitic magmas of the Jingpohu array (Fig. source(s). Our approach to constraint the degree 11b). McKenzie and Bickle’s calculation shows of partial melting to produce basanite is based on that more than 10Ð3% melt will separate from the a plot of La/Yb vs. Gd/Yb (Fig. 11). Frey et al. matrix if their viscosity is 0.1 Paás. According to (1978) have argued that primary melts in equilib- Shaw (1972)’s equation for viscosity, basanitic rium with a peridotite have Mg# values of 68 to melt is estimated to be 0.93 Paás. Thus, 0.5% 77. Thus, basanites DX4-8 and DX4-28 resemble basanitic melt may have separated from their

Fig. 11. (a) La/Yb vs. Gd/Yb variations as a function of degree of equilibrium melting of garnet lherzolites and garnet-amphibole-phologite-lherzolites (after Marzoli et al., 2000). Incompatible element depleted garnet lherzolite and enriched garnet-amphibole-phologite-lherzolite compositions are from Mckenzie and O’Nions (1991) and Ionov et al. (1996), respectively. Partition coefficients for olivine, pyroxenes and garnet are from McKenzie and O’Nion (1991), and those for amphibole and phologite are from Adam et al. (1993) and Adam and Green (1994). Modal and melting proportions for garnet lherzolite are from McKenzie and O’Nions. For garnet-amohibole- phologite lherzolite, modal proportions are: olivine 0.55; orthopyroxene 0.19; cliopyroxene 0.07; garnet 0.08; amphibole 0.06; phologite 0.05. (b) Primitive mantle normalized incompatible element concentrations of primary melts calculated from 90% garnet lherzolite (Gt-Lherz.) and 10% garnet-amphibole-phologite lherzolite (Gt-Am- Phol-Lherz.), compared with compositions of primitive alkali basalts. 150 Z. Zhang et al. source. In summary, low degrees of partial melt- from asthenosphere whereas melting of ing of a mixture of an incompatible element de- asthenosphere is related to high extension strength. pleted anhydrous lherzolite asthenospheric man- tle source and a fertile, amphibole-phlogopite-(and Acknowledgments—We thank the Directors of Bureau apatite-)bearing lherzolite continental lithospheric of Geology and Mineral Resources of Heilongjiang mantle source can account for the common chemi- Province and Heilongjiang Institute of Geological Sci- ences, for their support of this project. We are also cal signature of the Jingpohu Holocene basanites. grateful to other researchers, Heilongjiang Institute of Geological Sciences for providing assistance in collect- ing samples, logistical support and accommodation CONCLUSION during our staying in field. We appreciate helpful dis- The Jingpohu Holocene basaltic rocks consist cussions with Dr. Xu Yigang. Qiao Guangsheng, Chen Xiaoqing and Zhen Miaozi are thanked for their tech- of basanites, alkali olivine basalts and tephrites. nical and analytical collaboration. We acknowledge These magmas were derived from the upper man- constructive reviews of the paper by Dr. Yuji Orihashi tle, as indicated by megacrysts of kaersutite, and Dr. Morris. Financial support is from National Sci- phlogopite and anorthoclase, and by the presence ence Fundation of China (No. 40172026) and National of mantle xenoliths in tephrites and alkali olivine Key Fundamental Research Project (No. G1999043205). basalts. 6.8% olivine, 0.8% augite, 2.8% magnet- ite and 1.3% Cr-spinel were extracted from a rela- tively primitive basanite to produce the alkali REFERENCES olivine basalt, whereas the generation of the Adam, J. (1990) The geochemistry and experimental tephrite requires the fractionation of kaersutite petrology of sodic alkaline basalts from Oatlands, (15.5%), phlogopite (5.7%), anorthoclase (26.6%), Tasmania. J. Petrol. 31, 1201Ð1223. leucite (2.0%), magnetite (5.9%), olivine (3.0%) Adam, J. and Green, T. H. (1994) The effects of pres- and augite (6.4%) from the relatively primitive sure and temperature on the partitioning of Ti, Sr and basanite. The fractionation occurred in the upper REE between amphibole, clinopyroxene and 87 86 basanitic melts. Chem. Geol. 117, 219Ð234. mantle. Some samples with high Sr/ Sr and low Adam, J., Green, T. H. and Sie, S. H. (1993) Proton 143 144 Nd/ Nd may have undergone slight contami- microprobe determined partitioning of Rb, Sr, Ba, Y, nation en route to the surface by rocks of lower Zr, Nb and Ta between experimentally produced crust. Relatively unradiogenic isotopic ratios and and silicate melts. Chem. Geol. 109, 29Ð relatively high contents of highly incompatible 49. elements suggest that the source of the basanites Baker, D. K. and Willey, P. J. (1992) High pressure apatite solubility in carbonate-rich liquids: implica- is a mixture of a dry, depleted, lherzolite tions for mantle metasomatism. Geochim. asthenospheric mantle source and a fertile, Cosmochim. Acta 56, 3409Ð3422. amphibole-phologite-(and apatite-)bearing Barton, M. and Hamilton, A. (1982) Water- lherzolite continental lithospheric mantle source. undersaturated melting experiments bearing upon the From REE and trace element contents in the lavas, origin of potassium-rich magma. Miner. Mag. 45, 267Ð278. low degrees of partial melting of a mixture of Beccaluva, L., Di Girolamo, P. and Serri, G. (1991) about 90% depleted anhydrous lherzolite Petrogenesis and tectonic setting of the Roman vol- asthenospheric mantle source and 10% enriched, canic province, Italy. Lithos 26, 191Ð221. amphibole-phlogopite-(and apatite-)bearing Chapman, N. A. (1976) Inclusions and megacrysts from lherzolite continental lithospheric mantle source undersaturated tuffs and basanites, East Fife, Scot- land. J. Petrol. 17, 472Ð498. can account for the common chemical signature Chazot, G., Menzies, M. A. and Harte, B. (1996) De- of the Jingpohu Holocene basanites. Melting and termination of partition coefficients between apatite, enrichment of continental lithosphere are related clinopyroxene, amphibole, and melt in natural spinel to asthenosphere upwelling and metasomatism of lherzolites from Yemen: implications for wet melt- fluids very rich in potassium and volatiles released ing of the lithospheric mantle. Geochim. Cosmochim. The Jingpohu Holocene alkali basaltic rocks, northeastern China 151

Acta 60, 423Ð437. ation in space and time. J. Geophys. Res. 96, 13693Ð Chen, H. and Xu, Y. (1992) The tectonic and environ- 13711. ment of Cenozoic volcanic rocks and seism activity Foley, S. (1991) High-pressure stability of the fluor- in Heilongjiang province. Volcanism and and hydroxy-endmembers of pargasite and K- Humanbeings Environment (Liu Ruoxin, ed.), 366Ð richterite. Contrib. Mineral. Petrol. 55, 2689Ð2694. 391, Seism Publishing House (in Chinese). Forbers, W. C. (1974) Phase relations of Class, C. and Goldstein, S. L. (1997) Plume- titanphlogopite, K Mg Ti Al Si O (OH): a refractory lithospheric interactions in the ocean basins: con- phase in the upper mantle? Earth Planet. Sci. Lett. straints from the source mineralogy. Earth Planet. 22, 60Ð66. Sci. Lett. 150, 245Ð260. Frey, F. A., Green, D. H. and Roy, S. D. (1978) Inte- Comin-Chiarmonti, P., Cundari, A., Piccirillo, E. M., grated models of basalt petrogenesis: a study of Gomes, C. B., Castorina, F., Censi, P., De Min, A., quartz tholeiites to olivine melililites from South Speziale, S. and Velazquez, V. F. (1997) Potassic and Eastern Australia utilizing geochemical and experi- sodic igneous rocks from eastern Paraguay: their ori- mental petrological data. J. Petrol. 91, 463Ð513. gin from a lithospheric mantle and genetic relation- Gallagher, K. and Hawkesworth, C. J. (1992) Dehy- ships with the associated Parana flood tholeiites. J. dration melting and the generation of continental Petrol. 38, 495Ð528. flood basalts. Nature 358, 57Ð59. Cox, K. G. (1980) A model for flood basalt volcanism. Geist, D. J., Baker, B. H. and McBirney, A. R. (1985) J. Petrol. 21, 629Ð650. GPP, a program package for creating and using Davies, G. R. and Macdonald, R. (1987). Crustal in- geochemical data files. University of Oregon, fluences in the petrogenesis of the Naivasha basalt- Eugene. comendite complex: combined trace element and Sr- Green, D. H. (1973) Conditions of melting of basanite Nd-Pb isotopic constraints. J. Petrol. 28, 1009Ð1031. magma from garnet peridotite. Earth Planet. Sci. Lett. Dulski, P. (1994) Interferences of oxide, hydroxide and 17, 456Ð465. chloride analyses species in the determination of rare Green, D. H. and Fallon, T. J. (1998) Pyrolite: a earth elements in geological samples by inductively Ringwood concept and its current expression. The coupled plama-mass spectrometry. Fresenius’s Jour- Earth’s Mantle (Jackson, I., ed.), 311Ð378, Cam- nal of Analytical Chemistry 350, 194Ð203. bridge University Press. Edgar, A. D. (1987) The genesis of alkaline magmas Green, D. H. and Ringwood, A. E. (1976) The genesis with emphasis on their source regions: inferences of basaltic magmas. Contrib. Mineral. Petrol. 15, from experimental studies. Alkaline Igneous Rocks 103Ð124. (Fitton, J. G. and Upton, B. G. J., eds.), Geol. Soc. Gulson, B. L., Korsch, M. J., Vaasjoki, M., Mizon, K. Spec. Publ. 30, 29Ð52. J., Porritt, P. M., Carr, G. R., Kamper, C., Dean, J. A. Edgar, A. D., Condliffe, R. L., Barnett, R. J. and Shirran, and Calvez, J. Y. (1984) Lead isotope ratio measure- R. J. (1976) Experimental petrology of a highly po- ments using the Isomass 54E in fully automated tassic magma. J. Petrol. 17, 339Ð356. mode. Int. J. Mass Spectrom. Ion Proc. 59, 125Ð142. Fan, Q. C. and Hooper, P. R. (1991). The Cenozoic Haase, K. M. and Devey, C. W. (1994) The petrology basaltic rocks of eastern China: petrology and chemi- and geochemistry of Vesteris cal composition. J. Petrol. 32, 765Ð810. Basin—an intraplate alkaline volcano of non-plume Fancis, D. and Luden, J. (1995) The signature of origin. J. Petrol. 35, 295Ð328. amphibole in mafic alkaline lavas, a study in the Hamelin, B. and Allegre, C. J. (1985). Large-scale re- northern Canadian Cordillera. J. Petrol. 36, 1171Ð gional units in the depleted upper mantle revealed 1191. by an isotope study of the south-west Indian ridge. Faure, G. (1986) Principles of Isotope Geology. John Nature 315, 196Ð198. Wiley & Sons, 141Ð247. Harmer, R. H., Eglington, B. M., Farrow, D., Butcher, Faure, G. and Powell, J. C. (1972) Strontium Isotope A. R., Auret, J. M., Stander, Y. Y. and Grosser, E. Geology. Springer-Verlag, Berlin-Heidebery, New (1986) Manual of Laboratory Procedures for Isotope York, 171 pp. Analysis. Int. Rep. National Physical Research Labo- Fitton, J. G. and Dunlop, H. M. (1985) The Cameroon ratory, CSIR, Pretoria. Line, West Africa, and its bearing on the origin of Hart, S. R. (1984) The DUPAL anomaly: a large-scale oceanic and continental alkali basalt. Earth Planet. isotopic anomaly in the southern hemisphere. Nature Sci. Lett. 72, 23Ð28. 309, 753Ð756. Fitton, J. G., James, D. and Leeman, W. P. (1991) Ba- Hawkesworth, C. J., Kempton, P. D., Rogers, N. W., sic magmatism associated with Late Cenozoic exten- Ellam, R. M. and Calsteren, P. W. (1990) Continen- sion in the western United States: compositional vari- tal mantle lithosphere, and shallow level enrichment 152 Z. Zhang et al.

processes in the Earth’s mantle. Earth Planet. Sci. Mckenzie, D. (1989) Cratonic, circumeratonic and oce- Lett. 96, 256Ð268. anic mantle domians beneath the western United Heilongjiang Bureau of Geology and Mineral Re- states. J. Geophys. Res. 94, 7899Ð7915. sources (1991) Regional Geology of Heilongjiang Mckenzie, D. and Bickle, M. J. (1988) The volume and Province. Geological Publishing House, 569 pp. (in composition of melt generated by extension of the Chinese with English abstract). lithosphere. J. Petrol. 29, 53Ð72. Hirose, K. and Kushiro, I. (1993) Partial melting of dry Mckenzie, D. and O’Nions, R. K. (1991) Partial melt peridotites at high pressures: determination of com- distributions from inversion of rare earth element positions of melts segregated from peridotite using concentrations. J. Petrol. 32, 1021Ð1091. aggregates of diamond. Earth Planet. Sci. Lett. 114, Mckenzie, D. and O’Nions, R. K. (1995) The source 477Ð489. regions of ocean island basalts. J. Petrol. 36, 133Ð Hofmann, A. W. (1988) Chemical differentiation of the 160. Earth: the relationship between mantle, continental Mengel, K. and Green, D. H. (1989) Stability of crust, and oceanic crust. Earth Planet. Sci. Lett. 90, amphibole and phlogopite in metasomatized 297Ð314. peridotite under water-saturated and water- Hofmann, A. W. (1997) Mantle geochemistry: the mes- undersaturated conditions. Kimberlites and Related sage from oceanic volcanism. Nature 385, 219Ð229. Rocks (Ross, J., ed.), Geological Society of Australia, Hollister, L. S., Grissom, G. C., Peters, E. K., Stowell, Special Publication, 14, 571Ð581. H. H. and Sisson, V. B. (1987) Confirmation of the Menzies, M. A. (1990) Archaean, Proterozoic, and empirical correlation of Al in hornblende with pres- Phanerzoic lithospheres. Continental Mantle sure of solidification of calc-alkaline plutons. Am. (Menzies, M. A., ed.), 67Ð86, Oxford Science Publi- Mineral. 72, 231Ð237. cations. Hsu, C. and Chen, J. (1998) Geochemistry of late Menzies, M. A. and Hawkesworth, C. J. (eds.) (1987) Cenozoic basalts from Wudalianchi and Jingpohu Mantle Metasomatism. Academic Press, London. areas, Heilongjiang Province, northeast China. J. Mysen, B. O. and Kushiro, I. (1977) Compositional Asian Earth Sci. 16, 385Ð405. variations of coexisting phases with degree of melt- Ionov, D. A., O’Reilly, S. Y., Genshaft, Y. S. and ing of peridotite in the upper mantle. Amer. Miner. Kopylova, M. G. (1996) Carbonate-bearing mantle 62, 843Ð865. peridotite xenoliths from Spitsbergen: phase relation- Ormerod, D. S., Rogers, N. W. and Hawkesworth, C. J. ships, mineral compositions and trace element resi- (1991) Melting of the lithospheric mantle: inverse dence. Contrib. Miner. Petrol. 125, 375Ð392. modelling of alkali-olivine basalts from the Big Pine Irving, A. J. and Frey, F. A. (1984) Trace element abun- volcanic field, California. Contrib. Mineral. Petrol. dance in megacrysts and their host basalts: constraints 108, 305Ð317. on partition coefficients and megacryst genesis. Peng, Z. C., Zartman, R. E., Futa, E. and Chen, D. G. Geochim. Cosmochim. Acta 48, 1202Ð1221. (1986) Pb-, Sr- and Nd-isotopic systematics and Lee, D. C., Halliday, A. N., Davies, G. R., Essene, E. chemical characteristics of Cenozoic basalts, east- J., Fitton, J. G. and Temdjim, R. (1996) Melt enrich- ern China. Chem. Geol. 59, 3Ð33. ment of shallow depleted mantle: a detailed petro- Perry, F. V., Baldridage, W. S. and De paolo, D. J. (1987) logical, trace element and isotopic study of mantle- Role of asthenosphere and lithosphere in the genesis derived xenoliths and megacrysts from the Cameroon of Late Cenozoic basaltic rocks from the Rio Grande Line. J. Petrol. 38, 415Ð441. rift and adjacent regions of the southwestern United Liu, B., Chen, Y. and Zhu, B. (1989) Cenozoic basalts States. J. Geophys. Res. 92, 9193Ð9213. from Jinbohu, northeast China and chemical charac- Prestvik, T. and Goles, G. G. (1985) Comments on the teristics of their mantle source—Sr-Pb isotope and petrogenesis and the tectonic setting of Columbia trace element evidence. Geochimica 18, 9Ð19 (in River Basalts. Earth Planet. Sci. Lett. 72, 65Ð73. Chinese with English abstract). Ringwood, A. E. (1975) Composition and Petrology of Liu, C. Q., Masuda, A. and Xie, G. H. (1994) Major- the Earth’s Mantle. McGraw-Hill, New York, 618 pp. and trace-element compositions of Cenozoic basalts Rogers, N. W., Hawkeworth, C. J. and Ormerod, D. S. in eastern China: petrogenesis and mantle source. (1995) Late Cenozoic basaltic magmatism in the Chem. Geol. 114, 19Ð42. Western Great Basin, California and Navada. J. Marzoli, A., Piccirillo, E. M., Renne, P. R., Bellieni, Geophys. Res. 100, 10287Ð10303. G., Lacumin, M., Nyobe, J. B. and Tongwa, A. T. Sack, R. O., Walker, D. and Carmichael, I. S. E. (1987) (2000) The Cameron volcanic line revisited: Experimental petrology of alkali lavas: constraints petrogenesis of continental basaltic magmas from on cotectics of multiple saturation in natural basic lithospheric and AM sources. J. Petrol. 41, 87Ð109. liquids. Contrib. Mineral. Petrol. 96, 1Ð23. The Jingpohu Holocene alkali basaltic rocks, northeastern China 153

Shaw, H. R. (1972) Viscosities of magmatic silicate liq- end-member compositions: trace element and iso- uids: an empirical method of prediction. Am. J. Sci. topic constraints. Earth Planet. Sci. Lett. 104, 381Ð 272, 870Ð893. 397. Shen, Q., Xu, H., Zhang, Z., Gao, J., Wu, J. and Ji, C. Wenrich, K. J. (1995) Spatial migration and (1991) Early Precambrian Granulites in China. Geo- compositional changes of Miocene-Quaternary logical Press, Beijing. magmatism in Western Grand Canyon. J. Geophys. Sheng, Z., Wang, X. and Bai, Z. (1983) Deep-source Res. 100, 10417Ð10440. xenoliths from the Cenozoic basalts in the Jingpohu Wilson, M. (1989) Igneous Petrogenesis. Chapman and region and their evolution. Geology of Heilongjiang Hall, London. Region No. 1, 12Ð22 (in Chinese). Wilson, M. and Downes, H. (1991) Tertiary-Quaternary Simonetti, A., Goldstein, S. L., Schmidberger, S. S. and extension-related alkaline magmatism in western and Viladker, S. G. (1998) Geochemical and Nd, Pb, and Central Europe. J. Petrol. 32, 811Ð849. Sr isotope data from Deccan alkaline complexes— Wilson, M., Downes, H. and Cehria, J. M. (1995) Con- inferences for mantle sources and plume-lithosphere trasting fractionation trends in coexisting continen- interaction. J. Petrol. 39, 1847Ð1864. tal alkaline magma series: Catal, Massif Central, Stuiver, M. and Reimer, P. J. (1993) Radiocarbon cali- France. J. Petrol. 36, 1729Ð1753. bration program 3.0. Radiocarbon, 35, 215Ð230. Wu, F., Sun, D. and Lin, Q. (1999) Petrogenesis of the Sun, S. S. and McDonough, W. E. (1989) Chemical and Phanerozoic granites and crustal growth in North- isotopic systematics of oceanic basalts: implications east China. Acta Petrologica Sinica 15, 181Ð189 (in for mantle composition and processes. Magmatism Chinese with English abstract). in the Ocean Basins (Saunders, A. D. and Norry, M. Wulff-Pedersen, E., Neumann, E. R. and Jensen, B. B. J., eds.), Geological Society, London, Special Publi- (1996) The upper mantle under La Palma, Canary cations, 42, 313Ð346. Island: formation of Si-K-Na-rich melt and its im- Tatsumoto, M., Basu, A. R., Huang, W. K., Wang, J. portance as a metasomatic agent. Contrib. Mineral. W. and Xie, G. H. (1992). Sr, Nd, and Pb isotopes of Petrol. 125, 113Ð139. ultramafic xenoliths in volcanic rocks of Eastern Yoder, H. S. and Tilley, C. E. (1962) Origin of basalt China: enriched components EMI and EMII in sub- magmas: an experimental study of natural and syn- continental lithosphere. Earth Planet. Sci. Lett. 113, thetic rock systems. J. Petrol. 2, 342Ð532. 107Ð128. Zartman, R. E. (1984) Lead, strontium,and neodym- Taylor, S. R. and McLennan, S. M. (1995) The ium isotopic characterization of mineral deposits rela- geochemical evolution of the continental crust. Rev. tive to their geologic environments. Proc. 27th Int. Geophys. 33, 241Ð265. Geol. Congr., Moscow, VNUK Science Press, Thompson, R. N., Morrison, M. A., Hendry, G. L. and Utrecht, Vol. 12, 83Ð106. Parry, S. J. (1984) An assessment of the relatively Zhang Z., Li, Z., Li, S., Xin, Y., Li, Z. M. and Wang, roles of crust and mantle magma genesis: an elemen- X. (1999) First discovery of kaersutite from the tal approach. Phil. Trans. R. Soc. Lond. A310, 549Ð Holocene basalts in the Jingpohu area, Heilongjiang 590. province, and its significance. Geol. Rev. 45(5), 486 Tu, K., Xie, G., Zhang, M., Wang, J., Flower, M. F. J. (in Chinese with English abstract). and Carlson, R. W. (1992) Sr, Nd, Pb isotopic com- Zhang, Z., Li, Z., Li, S., Xin, Y., Li, Z. M. and Wang, ponent of Cenozoic basalt in eastern China—Isotopic X. (2001) Discovery of a special type of leucite evidence on sub-continental lithosphere. Chronology tephrite in the Jingpohu area of Heilongjiang Prov- and Geochemistry of Cenozoic Volcanic Rocks in ince and its geological significance. Scienta China (Liu, R., ed.), 330Ð338, Seism Publishing Geologica Sinica 36, 23Ð35 (in Chinese with Eng- House (in Chinese). lish abstract). Viereck, L. G., Flower, M. F. J., Hertogen, J., Zhao, C., Peng, Y., Dang, Z. and Zhang, Y. (1996) Tec- Schminche, H. U. S. and Jenner, G. A. (1989) The tonic Framework and Crust Evolution of Eastern Jilin genesis and significance of N-MORB sub-types. and Heilongjiang Province. Liaoning University Contrib. Mineral. Petrol. 102, 112Ð126. Press (in Chinese). Wass, S. Y. (1979) Multiple origins of clinopyroxenes Zhou, X. and Zhu, B. (1992) Cenozoic isotopic system in alkali basaltic rocks. Lithos 12, 115Ð132. and division of mantle chemistry in eastern China. Watson, E. B. (1980) Apatite and phosphorus in man- Chronology and Geochemistry of Cenozoic Volcanic tle source regions: an experiment study of apatite/ Rocks in China (Liu, R., ed.), 366Ð391, Seism Pub- melt equilibria at pressures to 25 kbar. Earth Planet. lishing House, Beijing (in Chinese). Sci. Lett. 51, 322Ð335. Zindler, A. and Hart, S. R. (1986) Chemical dynamics. Weaver, B. L. (1991) The origin of ocean island basalt Ann. Rev. Earth Planet. Sci. Lett. 14, 493Ð571.