Earth and Planetary Science Letters, 69 (1984) 141-158 Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands
Geochemistry of tholeiitic and alkalic lavas from the Koolau Range, Oahu, Hawaii: implications for Hawaiian volcanism Michael F. Roden 1,,, Frederick A. Frey 1 and David A. Clague 2 I Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139 (U.S.A.) 2 United States Geological Survey, Menlo Park, CA 94025 (U.S.A.)
Received August 5, 1983 Revised version received March 23, 1984
Lavas of the post-erosional, alkalic Honolulu Volcanics have significantly lower S7Sr/S6Sr and higher 143Nd/l~Nd than the older and underlying Koolau tholeiites which form the Koolau shield of eastern Oahu, Hawaii. Despite significant compositional variation within lavas forming the Honolulu Volcanics, these lavas are isotopically (Sr, Nd, Pb) very similar which contrasts with the isotopic heterogeneity of the Koolau tholeiites. Among Hawaiian tholeiitic suites, the Koolau lavas are geochemicallydistinct because of their lower iron contents and Sr and Nd isotopic ratios which range to bulk earth values. These geochemical data preclude simple models such as derivation of the Honolulu Volcanics and Koolau tholeiites from a common source by different degrees of melting or by mixing of two geochemicallydistinct sources. There may be no genetic relationship between the origin and evolution of these two lava suites; however, the trend shown by Koolau Range lavas of increasing 143Nd/144Nd and decreasing 87Sr/86Sr with decreasing eruption age and increasing alkalinity also occurs at Haleakala, East Molokai and Kauai volcanoes. A complex mixing model proposed for Haleakala lavas can account for the variations in Sr and Nd isotopic ratios and incompatible element abundances found in lavas from the Koolau Range. This model may reflect mixing and melting processes occurring during ascent of relatively enriched mantle through relatively depleted MORB-related lithosphere. Although two isotopically distinct components may be sufficient to explain Sr and Nd isotopic variations at individual Hawaiian volcanoes, more than two isotopically distinct materials are required to explain variations of Sr, Nd and Pb isotopic ratios in all Hawaiian lavas.
1. Introduction The d e v e l o p m e n t of H a w a i i a n volcanoes is characterized by distinct eruptive stages ['1]: the m a i n shield is d o m i n a n t l y composed of tholeiitic basalt b u t smaller volumes of alkalic basalt may e r u p t early a n d late d u r i n g the shield-building stage; the final phase of volcanic activity, the post-erosional stage, occurs after a period of
This paper has not been reviewed for standards of classification and nomenclature adopted by the U.S. Geological Survey. * Now at Department of Geology, University of Georgia, Athens, GA 30602, U.S.A. 0012-821X/84/$03.00
© 1984 Elsevier Science Publishers B.V.
volcanic quiescence a n d is formed of alkalic lavas which erupt from vents scattered over the shield. A t four H a w a i i a n volcanoes studied in detail ( K a u a i ; K o o l a u , O a h u ; East M o l o k a i a n d Haleakala, East Maui), the alkalic, post-erosional lavas have lower 87Sr/86Sr, higher 143Nd/144Nd a n d higher i n c o m p a t i b l e element a b u n d a n c e s t h a n their respective u n d e r l y i n g shield tholeiites [2-13]. A p p a r e n t l y , such trends are characteristic of H a w a i i a n volcanism, a n d they must reflect imp o r t a n t characteristics of how " h o t s p o t " volcanism evolves at i n d i v i d u a l volcanoes. I n this p a p e r we use geochemical data to evaluate the petrogenetic relationship between the tholeiitic lavas f o r m i n g the K o o l a u shield a n d lavas f o r m i n g
J 158° 00'
[58 ° 00'
-I ......... 50'
~,65 PAL-2 PAL
. . . .
SYMBOLS ALKALIC LAVAS (~ THOLEIITES
[ I/2 0 I 2 3 4 MILES I 5 0 rP 2I 5I 4I 5I KILOMETERS L I J
OVERLAY FOR TOPOGRAPHIC MAP OF THE ISLAND OF OABU, HAWAII SCALE h 62500
Z ~°lSl 157°40,
Fig. 1. Map of Oahu showing locations of Koolau Range samples analyzed for trace elements and isotopic ratios of Sr and Nd. Tholeiitic samples ( X within open circle) are from diverse portions of the shield; however, all vents and samples of the Honolulu Volcanics (solid circle within open circle) are from the southeast part of the Koolau Range.
158 ° 15'
143 the overlying post-erosional Honolulu Volcanics. These samples are particularly appropriate for study because their Sr and Nd isotopic ratios as well as major and trace element abundances nearly span the ranges found in Hawaiian lavas (e.g. [2-13]). In addition, Clague and Frey  presented a detailed model for the source of the Honolulu Volcanics, but they did not consider the Koolau tholeiites and Stille et al.  presented isotopic data for Koolau tholeiites and the Honolulu Volcanics, but they did not utilize elemental abundances to constrain their petrogenetic models. The Koolau tholeiitic shield forms the eastern part of Oahu, Hawaii (Fig. 1) and it has a subaerial plus submarine volume of - 20.9 x 103 km 3 . Subaerial Koolau tholeiites have K-Ar ages of 1.8 2.6 m.y. . Conformable, alkalic-capping stage lavas are absent, but the post-erosional Honolulu Volcanics (< 1 m.y., ) comprise 37 distinct groups of vents which occur along linear zones perpendicular to the main rift system of the underlying Koolau shield . Relative to the tholeiitic shield, lavas from these vents are volumetrically insignificant ( < 1% of the shield), and they comprise a series of alkalic basalt, nephelinite and melilitite. Lavas of the Koolau shield were the subject of several classic studies [16-18], but recent emphasis [3, 6-8] has been on the Honolulu Volcanics, in large part because they contain abundant mantle xenoliths. On the basis of a detailed compositional study of the Honolulu Volcanics, Clague and Frey  concluded that much of the compositional diversity among the Honolulu Volcanics can be explained by different degrees of melting of a compositionally homogeneous source. Sr, Nd, Pb and Hf isotopic differences between Koolau tholeiites and the Honolulu Volcanics are well-documented [2,4,6, 7,10,19]. Our objectives are to present additional geochemical data which aid in defining differences between the sources of the Koolau tholeiites and the Honolulu Volcanics and to discuss models consistent with the geochemical trends established by these lavas forming the Koolau Range.
2. Sample description Clague and Frey  divided the Honolulu Volcanics into three lava types: (a) alkali olivine basalt and basanite, (b) nephelinite and (c) melilitite. Within the last two groups, they distinguished high-TiO 2 from low-TiO 2 lavas. For Sr and Nd isotopic studies, we selected alkali basalt (65KAL1), basanite (65PAL-2), four nephelinites, three in the low-TiO 2 group (66PY-1, 68PB-2, 69WIL-1) and one in the high-TiO 2 group (65KAPAA-11), and two melilitites (MQ2-L, PEG) of the high TiO 2 group. All samples except the melilitites are splits of samples studied by Clague and Frey  who presented major and trace element abundance data for these rocks. The melilitites are from the flow exposed in Moiliili Quarry . PEG is a sample of a felsic, pegmatoidal segregation in the host melilitite, MQ2-L. This host rock is petrographically and compositionally similar to 65MOIL-2 studied by Clague and Frey . Sample locations are shown in Fig. 1. Twelve Koolau tholeiites were analyzed for major elements. Sample locations (Fig. 1) and brief petrographic descriptions are in Appendix 1. Although it is difficult to obtain fresh Koolau lavas, eight of these samples contain < 0.5% H2 O÷ and have Fe203/FeO < 0.56; five of this group were selected for isotopic and trace element studies. In addition, major and trace element abundances and 87Sr/86Sr ratios are available for four additional Koolau samples [3,6,21], and we determined 143Nd/144Nd in one of these samples (65 MAKI-1). Rare earth element (REE) abundances were previously determined in the most MgO-rich (11.7%) sample, WW9948 .
3. Analytical methods Major element compositions (Tale 1) were determined by classical wet chemical techniques by the U.S. Geological Survey (Denver, Colo.); Rb and Sr were determined by isotope dilution at M.I.T. , Ba, V, Ni, Cu, Zn, Ga, Y, Zr and Nb were determined by X-ray fluorescence at the University of Massachusetts  and Sc, Cr, Co, Hf, REE, Ta and Th by instrumental neutron activa-
Less O Total
51.74 15.29 3.91 7.00 5.59 9.73 2.84 0.26 0.38 0.41 2.20 0.28 0.15 0.00 nd nd 0.01 0.03 nd 99.82 0.00 ~
48.23 14.45 3.46 7.38 6.80 9.98 2.06 1.11 2.05 1.81 2.10 0.25 0.17 0.03 nd nd 0.00 0.032 0.020 99.93
D102543 D 1 0 2 7 8 5 68TAN-2 7OULUP-8
51.17 13.81 4.70 5.76 6.28 9.62 2.50 0.52 1.19 1.37 2.36 0.33 0.14 0.02 0.01 0.04 nd nd nd 99.82
51.83 14.48 4.22 5.87 6.02 9.30 2.62 0.71 1.79 0.38 2.23 0.30 0.15 0.01 0.01 0.04 nd nd nd 99.96
52.50 14.40 2.71 7.90 6.42 9.60 2.44 0.52 0.40 0.21 2.23 0.23 0.16 0.01 0.01 0.04 nd nd nd 99.83
51.27 13.10 2.49 8.02 11.68 8.16 2.51 0.14 0.31 0.34 1.54 0.20 0.16 0.01 0.13 0.02 nd nd nd 100.08
51.81 13.06 2.64 7.78 9.55 8.68 2.47 0.56 0.47 0.35 1.94 0.27 0.15 0.01 0.02 0.04 nd nd nd 99.80
51.08 14.59 3.44 7.89 6.74 10.08 2.63 0.16 0.06 0.32 2.17 0.25 0.17 0.01 0.01 0.03 nd nd nd 99.63 0.01 99ff7.97
54.57 14.05 2.69 7.49 6.86 8.11 3.02 0.40 0.23 0.59 1.74 0.23 0.15 0.01 0.01 0.03 nd nd nd 99.98
D102240 D102241 WW9991 WW10398
51.99 14.28 2.37 8.38 6.94 9.56 2.69 0.37 0.43 0.24 2.21 0.29 0.16 0.00 0.02 0.03 nd nd nd 99.96
48.80 18.88 4.83 6.00 3.73 9.15 4.06 0.71 0.04 0.15 2.83 0.38 0.13 0.00 0.01 0.04 nd nd nd 99.74
All analyses by wet chemistry in U.S.G.S. laboratories in Denver. Project Leader, L. Peck. Analysts are: Edythe E. Engleman (69TAN-2, 68KAV-6, 66KAPAA-16B, and 68HUL-2), Vertie C. Smith (WW9948, WW9980, WW9991, WW10398, WW10403, WWl1320, 69MAK-2 and 7OULUP-8).
52.76 14.32 3.46 6.78 6.62 9.47 2.71 0.50 0.33 0.24 2.05 0.28 0.17 0.02 nd nd 0.07 0.029 0.017 99.83
SiO2 A1203 Fe203 FeO MgO CaO Na20 K20 H2 O÷ H20TiO 2 P205 MnO CO 2 CI F S Cr203 NiO Subtotal
Chemical analyses of Koolau tholeiites
tion analysis at M.I.T. . Precision at the 95% confidence level is estimated at 2% for Rb and 1% for Sr. Precision and accuracy of the X-ray fluorescence technique were discussed in Rhodes . Precision of the neutron activation analyses was discussed in Ila and Frey  and Frey and Clague  and accuracy can be evaluated from data for BHVO-1 in Table 2. Isotopic ratios for Sr and Nd were determined by standard techniques [27,28]. Quoted errors refer to 95% confidence limits as calculated from within-run statistics; however, based on duplicate analyses these confidence limits are reasonable estimates of reproducibility (Tables 3 and 4). All isotopic ratios were normalized to 86Sr/SSSr= 0.1194 and 146Nd/144Nd=0.7219 and are reported relative to Eimer and Amend SrCO 3 87Sr//86Sr=0.70800 and BCR-1 ] 4 3 N d / / l n 4 N d = 0.51264.
4. Results 4.1. Honolulu Volcanics
The major and trace element geochemistry of the Honolulu Volcanics were discussed by Clague and Frey . A significant geochemical feature of the Honolulu Volcanics is the 4- to 5-fold variation in abundances of highly incompatible elements such as P, La, Ce and Th. The rare earth element (REE) abundances in the eight Honolulu Volcanics samples analyzed for Sr and Nd isotopes are indicated in Fig. 2. Sample 65KAL-1 has the lowest REE content and the nephelinites have REE contents near the maximum found by Clague and Frey . Sample MQ2-L is also very enriched in light REE (LREE) but like other samples from Moiliili Quarry it has unusually low heavy REE (HREE) abundances (Table 2, ). The pegmatitic
,,F,O,R 5o . . . . . .
AH I LO Ce
I I I Pm Sm Eu
RARE E A R T H ATOMIC NUMBER
Pm Sm Eu
RARE EARTH ATOMIC NUMBER
Fig. 2. Chondrite-normalized  REE abundances for samples analyzed for Sr and Nd isotopic ratios (Tables 3 and 4). A. 65KAL-1 (filled triangles), 65PAL-2 (open circles), 65KAPAA-11 (open triangles), 68PB-2 (filled circles), 66PY1 ( x ) , 69WIL-1 ( + ) , MQ2L (open squares), PEG (filled squares). These are new data obtained at M.I.T. (available from F. Frey) which are more precise and accurate than those obtained at the U.S.G.S. . Except for the pegmatitic sample (PEG) from vent 37 (Moiliili quarry) all samples lie within the range (indicated by diagonal lines) for 30 samples of the Honolulu Volcanics . However, note that MQ2L from Moiliili quarry has the lowest HREE abundances; low HREE abundances are characteristic of this flow . B. WM9948 (filled squares), 69MAK-2 (filled circles), WW10403 (+), WW9991 (×), WW9980 (open circles), 69TAN-2 (*). Four samples (69MAK-2, WW10403, WW9991 and WW9980) have similar REE abundances and an average of three other Koolau tholeiites  lies within the range of these four samples. Note that HREE abundances in these Koolau tholeiites overlap with those in the Honolulu Volcanics but that LREE abundances are significantly less in the Koolau tholeiites. Lower REE content of WW9948 reflect its relatively higher MgO content (11.7%, Table 1). The unusually high REE abundances and negative Ce anomaly in 69TAN-2 do not correlate with other geochemical parameters (Tables 1 and 2) or petrographic characteristics.
4.3 2.1 2.2
160 106 111 21
1.98 0.86 0.87
490 416 109 19
1.89 0.78 1.0
260 87 110 20
1.92 0.85 1.1
293 112 105 21
4,8 1.55 -
337 160 209
8.2 2.90 -
5 27 186
2.12 0.88 0.95
2.07 1.0 -
e B H V O - 1 analyses d o n e at M.I.T. c o n c u r r e n t l y with Loihi s e a m o u n t analyses . f BHVO-1 analyses by isotope dilution. A v e r a g e of d a t a f r o m Sun et al.  a n d Feigenson et al. .
d B H V O - 1 analysis d o n e at M.I.T. c o n c u r r e n t l y with K o o l a u samples.
a T r a c e e l e m e n t d a t a for four o t h e r K o o l a u tholeiites studied b y J a c k s o n a n d W r i g h t  were r e p o r t e d by B u d a h n  a n d three of these were used to c o m p i l e the K o o l a u a v e r a g e of L e e m a n et al. . R E E analyses of four o t h e r K o o l a u tholeiites w e r e r e p o r t e d by Schilling a n d W i n c h e s t e r . b G e o c h e m i c a l d a t a for o t h e r H o n o l u l u Volcanics a n a l y z e d for Sr a n d N d isotopes are in C l a g u e a n d F r e y . Wilkinson a n d Stolz  discuss the vent 37 flow in detail. c D a t a r e p o r t e d in F r e y [22, table 2].
1.87 0.87 1.1
Eu Tb Ho
230 109 103 20
Cr Ni Zn Ga
WW9948 c MQ2L
U.S.G.S. s t a n d a r d BHVO-1
H o n o l u l u Volcanics WW9980
( v e n t 37) b
K o o l a u tholeiites a
T r a c e element a b u n d a n c e s ( p p m ) in K o o l a u tholeiites a n d H o n o l u l u Volcanics f r o m vent 37 (Moiliili q u a r r y )
PAL-2, is analytically distinguishable from the mean for the data set. Previous 87Sr//86Sr analyses of lavas from the Honolulu Volcanics range from 0.70274 to 0.70354 [2,4-6,10]. The most recent detailed Sr isotopic studies of the Honolulu Volcanics yielded a weighted mean of 0.70340 + 0.00004 for 14 samples  and a mean of 0.70330 for 6 samples  (adjusted to E and A standard = 0.70800). Because of the small range in 87Sr/86Sr found in our study and these recent studies, we suspect that much of the reported 875r//86Sr variation in lavas from the Honolulu Volcanics is the consequence of analytical uncertainty and interlaboratory bias. When combined with the previously reported homogeneity of Pb and Hf isotopic ratios in the Honolulu Volcanics [10,19], these data establish the remarkable isotopic homogeneity of the wide variety of alkalic lavas forming the Honolulu Volcanics, and support the conclusions of Clague and Frey  that much of the compositional heterogeneity of these lavas results from differences in degree of melting of a compositionally uniform
sample (PEG) from this quarry has the highest REE abundances found in the Honolulu Volcanics (Fig. 2). Although the eight samples that we studied are representative of the diverse lava compositions occurring in the Honolulu Volcanics, these samples are nearly isotopically homogeneous (Table 3, Fig. 3). For example, all the 143Nd/144Nd ratios are within 2 x 10 -5 of the mean 143Nd/144Nd ratio, 0.513035, and individual 143Nd/la4Nd ratios are analytically indistinguishable from the mean. Moreover, 143Nd//144Nd ratios in six additional samples of the Honolulu Volcanics determined by Stille et al.  have a mean of 0.513067 (adjusted to BCR-1 = 0.51264), and these individual ratios are within 4.5 × 10-5 of our mean value (Fig. 3). Also, single 143Nd/laaNd analyses previously reported for samples from Moiliili quarry [4,5] are within analytical error of our lowermost 143Nd/144Nd ratio (Fig. 3). The six 87Sr/86Sr ratios reported here average 0.70334 and range from 0.70326 to 0.70339 (Table 3, Fig. 3). Only one analysis, that of basanite 65 TABLE 3 Sr and Nd isotopic ratios in the Honolulu Volcanics a Rb (ppm) Alkali olivine basalt: 65KAL-1 Basanite: 65PAL-2 Nephelinites: 66PY-1
0.70334 + 3
0.513024 + 25
0.70326 + 3
0.513048 + 25
0.70331 + 9 (0.70354 + 8)
0.513060 ___22 0.513027 + 26 0.513029 + 24 0.513017 + 17 0.513020+ 17 0.513031 5:26
0.70337 ___6 0.70336 +__8
0.513056 + 20 0.513027 5:21
Melilitites: MQ2-L PEG
0.70339 (0.70335 (0.70345 (0.70339
+ + + +
10 11) 10) 13)
a Rb and Sr abundances by isotope dilution. Sm and Nd abundances are results of INAA analyses done at M.I.T.; these data are more precise and supersede those of Clague and Frey  which were obtained in a U.S.G.S. laboratory; new data are available from Frey (M.I.T.). Analytical procedures are discussed in the text. Plus and minus values in isotopic ratios reflects 2 sigma uncertainty in last digits. Additional Sr isotopic data for Honolulu volcanics are in Lanphere and Dalrymple  who summarize earlier isotopic studies of the Honolulu Volcanics. Our Sr isotopic data agree well with the results of Lanphere and Dalrymple. Other recent Nd isotopic data for Honolulu Volcanics are in DePaolo and Wasserburg  (sample USNM 113095-50), O'Nions et al.  (sample HA-8) and Stille et al.  (6 samples). b Rb and Sr abundances and Sr isotopic ratios within parentheses are from Lanphere and Dalrymple . These isotopic ratios have been adjusted to E and A standard = 0.70800 (M. Lanphere, personal communication, 1983).
148 0.5132 T ~ - - - I
' M/R8 'FIELDE[D~ 1
0 513 ] ~
~ , ~ (typicol _+2 sigmo)
•i . . . . . . ....-,-t
0.7034 0,7036 07038 07040 0,7042 0,7044 0,7046 87S r /86Sr
Fig. 3. 143Nd/144Nd versus 875r//S6Sr plot showing partial fields for M O R B and bulk earth (the latter represented by the rectangle in lower right, 87Sr//86Sr = 0.7047 + 7 and 1 4 3 N d / l ~ N d = 0 . 5 1 2 6 2 ± 1 2 , from Zindler ). Hawaiian field defined by results from references 4, 5, 9, 10, 44 and 45. Fields for Koolau tholeiites and aikalic post-erosional Honolulu Volcanics are defined by data from this paper (filled circles and squares) and Stille et al.  (open circles and squares). Filled circles with plus sign are for Honolulu Volcanics (vent 37) reported in DePaolo and Wasserburg  and O'Nions et al. . Open circle with plus sign is for a Koolau tholeiite from DePaolo and Wasserburg . The isotopic heterogeneity of Koolau tholeiites is further established by 4 other samples with S7Sr/86Sr less than 0.70398 . All data plotted is relative to E and A standard 87sr/a6sr = 0.70800 and BCR-1 143Nd/144Nd = 0.51264.
source. Furthermore, samples with anomalous compositions, such as members of the low- and high-TiO 2 groups and samples with anomalously low HREE and Sc contents (vent 37 samples such as MQ2-L) are isotopically indistinguishable from other samples of the Honolulu Volcanics. 4.2. Koolau tholeiites
Historic tholeiitic flows from Mauna Loa and Kilauea can be distinguished by small but significant differences in major and trace element composition (e.g. [29-31]), and they differ in Pb isotopic ratios . Apparently recent magmas from these two volcanoes evolved from compositionally distinct parental magmas. When compared at a common MgO content (e.g., 7%), the Koolau tholeiites have SiO 2 and TiO 2 contents similar to
historic lavas from Mauna Loa, but Koolau tholeiites have lower total FeO contents than most Hawaiian tholeiites (Fig. 4 and [33, table 11]. Like other Hawaiian tholeiites, Koolau tholeiites range widely in MgO content and they have low A1203 contents (generally < 15%) relative to mid-ocean ridge basalts. The most MgO-rich samples, WW9948 and WW9980, have the highest Cr and Ni contents (Tables 1 and 2). Abundances of the compatible trace elements, Sc, V, Cr, Ni and Zn are typical for Hawaiian tholeiites (cf. Table 2 and
[31,34]). Relative to Kilauea tholeiites, abundances of incompatible trace elements, such as Sr, Nb, Hf and Th, are lower in Mauna Loa tholeiites [24,31,34]. Abundances of Nb, Hf and Th in Koolau tholeiites are intermediate between Kilauea and Mauna Loa tholeiites; however, Z r / N b ratios of Koolau tholeiites resemble Mauna Loa tholeiites, whereas Sr abundances in Koolau tholeiites resemble Kilauea tholeiites (cf. Table 2 and [24,31,34]). In the six Koolau samples studied in detail, the largest abundance variations are for K 2 0 and Rb (factors of 4 and 10, respectively).
*0 i IO I' ~
T ~ /
I0 12 MgO (wt. %)
Fig. 4. Total iron as FeO (FeO*) versus MgO illustrating that individual Hawaiian tholeii::c suites form sub-horizontal trends with nearly uniform FeO* contents. Fields indicated for Koolau tholeiites (this paper and ), and the dominantly tholeiitic H o n o m a n u suite from Haleakala . Mauna Loa trend defined by 33 samples (6.85-23.6% MgO) of historic age ( H I S T M L from Wright [29, table 3]). Kilauea trend defined by 20 samples (7.2-19.5% MgO) from 1959 eruption from Kilauea Iki crater (KILl959 from Wright [29, table 31). Koolau tholeiites have the lowest average FeO* among Hawaiian tholeiitic suites whereas the Honomanu series has one of the highest average FeO*. Above MgO = 7%, the sub-horizontal trends for individual s~aites can be explained by olivine fractionation and accumulation; however, the different FeO* content of individual suites probably reflects FeO* abundance differences in the parental magmas .
149 Although K 2 0 and Rb contents are positively correlated in K o o l a u lavas, abundances of other incompatible elements are not correlated with K 20 and Rb contents. It is well established that Hawaiian tholeiites have chondrite-normalized R E E patterns which are distinct from those of M O R B and most continental basalt [34,35]. The Koolau tholeiites that we studied (except 69TAN-2) have chondrite-normalized R E E patterns typical of Hawaiian tholeiites (Fig. 2), and their R E E abundances (Table 2) are similar to the average of 3 Koolau tholeiites reported by Leeman et al. . These Koolau tholeiites have L a / C e , L a / S m and L a / Y b ratios similar to those in Kilauea tholeiites. The MgO-rich (11.7%) sample WW9948 has a chondrite-normalized R E E pattern similar to other Koolau tholeiites but it has lower R E E contents; in contrast, tholeiite 6 9 T A N - 2 has distinctly higher R E E abundances than the other tholeiites, and it has a p r o n o u n c e d negative Ce anomaly (Fig. 2). The 2to 3-fold enrichment in R E E is accompanied by a similar increase in Y, but abundances of all other incompatible elements analyzed in 6 9 T A N - 2 are within 25% of those in other Koolau tholeiites (Table 2). Hence, 6 9 T A N - 2 is anomalously enriched in R E E and Y relative to other Koolau tholeiites and Hawaiian tholeiites in general. Our 87Sr/86Sr data (Table 4, Fig. 3) extend the
range previously reported for Koolau tholeiites, 0.70367-0.70439 [2,4,6,10], to higher 87Sr/86Sr ratios. Previously determined 143Nd/144Nd ratios (relative to BCR-1 = 0.51264) lie within the range of our data [4,10]. Ratios of 87Sr/86Sr and 143Nd/144Nd in Koolau lavas are inversely correlated and extend the field for Hawaiian lavas to near bulk-earth estimates for 87Sr/86Sr and 143Nd/ln4Nd (Fig. 3). A n important point is that our data and those of Lanphere and Dalrymple  establish that the Koolau tholeiites are isotopically heterogeneous. In contrast, the five samples analyzed by Stille et al.  are nearly homogeneous with 143Nd/144Nd from 0.51272 to 0.51275 and 87Sr/86Sr from 0.70408 to 0.70420 (relative to BCR-1 = 0.51264 and Eimer and A m e n d = 0.70800). In fact, 8 of 11 samples from this paper and Stille et al.  have Sr and N d isotopic ratios within narrow ranges, 87Sr/86Sr = 0.70408 to 0.70437 and 143Nd/144Nd = 0.51272 to 0.51276., However, 5 of the 8 samples analyzed by Lanphere and Dalrymple  have 87Sr/868r < 0.70398 (relative to Eimer and A m e n d = 0.70800). Because these three laboratories (M.I.T. and U.S.G.S. in Denver and Menlo Park) obtained similar Sr and N d isotopic ratios for the Honolulu Volcanics, and Koolau lavas from the lower section of M a k a p u u Point, we interpret the Koolau isotopic data as reflecting significant iso-
TABLE 4 Sr and Nd isotopic ratios in Koolau tholeiites a WW9991 WWI0403 69MAK-2 WW9980
Rb (ppm) 1.52 5.61 15.6 8.32
Sr (l~pm) 395 435 421 450
875r/ 86Sr 0.70380 + 3 0.70423 -+4 0.70427 + 7 0.70432 + 3 0.70441 +6 0.70452 -+6 0.70458 -+4 (0.70367 -+15)
Sm (ppm) 5.1 5.39 5.1 5.50 11.4
18.4 20.4 19.2 21.3
0.512880 + 20 0.512764 _+19 0.512750 + 21 0.5127245:21 0.512750-+ 13 0.512673 -+13
0.512810 -+36 0.512772+17
Rb and Sr abundances determined by isotope dilution. Sm and Nd abundances determined by neutron activation. Analytical procedures discussed in text. Plus and minus values on isotopic ratios refer to 2 sigma uncertainty in last digit (Sr) or last two digits (Nd). Sr isotopic data for Koolau tholeiites were reported by Powell and Delong  who analyzed 3 of the same samples. More recent Koolau tholeiite isotopic analyses are reported in DePaolo and Wasserburg  (sample USNM 113095-60), Lanphere and Dalrymple  (8 samples) and Stille et al. [101 (5 samples). b Data within parentheses for 65 MAKI-1 from Lanphere and Dalrymple  withS7Sr/86Sradjusted to E and A standard = 0.70800.
150 topic heterogeneity among the Koolau lavas. Because variations in Sr and Nd isotopic ratios correlate with stratigraphy of shield lavas at other Hawaiian volcanoes (Haleakala, ; Kohala, [43,47]; and East Molokai, [12,13]), it is important to establish if a similar correlation exists among Koolau lavas. Unfortunately, the stratigraphic relationships of most Koolau samples analyzed for Sr and Nd isotopes are not known. In the Makapuu Point section the uppermost flow has a lower 87Sr/86Sr ratio (0.70376, ) than flows (5 samples) from lower in the section (> 0.70408, this paper and [6,10]). This is the same correlation, decreasing 87Sr/86Sr with decreasing age, found at other Hawaiian volcanoes, but definitive evaluation of isotopic variations with eruption age of Koolau lavas requires study of many samples with known stratigraphy (Frey et al., in progress). Stille et al.  noted "that Koolau Sr, Nd, Hf and Pb isotopic compositions are close to presumed "bulk Earth" values, yet do not plot on the mantle p l a n e . . . " proposed by Zindler et al.  to describe the Pb, Sr and Nd isotopic variations of oceanic basalts. Our confirmation of isotopic heterogeneity among Koolau lavas and the scatter of Koolau points in Fig. 3 indicate that realistic average isotopic ratios for the Koolau shield will be difficult to obtain. However, it is possible that the average will lie closer to the mantle plane than the data of Stille et al. . Because the mantle plane was constructed from averages for individual (or closely related) oceanic islands and ridge segments, it may be more meaningful to plot the Koolau average when a more realistic estimate becomes available. 4. 3. Comparison of Honolulu Volcanics with Koolau tholeiites
In summary, the well established major element abundance differences between Koolau tholeiites and lavas of the Honolulu Volcanics are accompanied by significant differences in trace element contents. In particular, the alkalic lavas of the Honolulu Volcanics have higher contents of highly to moderately incompatible elements such as light REE, Nb, Hf, Ta, Th, Rb, Ba and Sr and higher
contents of compatible elements such as MgO, Ni and Cr (cf. Table 2 with tables 2 and 3 of Clague and Frey ). Also, abundance ratios of strongly incompatible trace elements to moderately incompatible trace elements (e.g., La/Sm, La/Ce, N b / Z r and Th/Sr) are higher in lavas of the Honolulu Volcanics than in Koolau tholeiites (Fig. 2 and Table 2). These compositional differences are accompanied by isotopic differences; for example, the Koolau tholeiites have higher 87Sr/86Sr and lower 143Nd/lnaNd than the Honolulu Volcanics (Fig. 3). Lavas of the Honolulu Volcanics are nearly isotopically homogeneous and overlap with the high 87Sr/86Sr, low 143Nd/laaNd end of the MORB field; in contrast, the Koolau tholeiites are isotopically heterogeneous and range to bulk earth estimates for 87Sr/86Sr and 143Nd/144Nd ratios (Fig. 3). In fact, Sr and Nd isotopic variations in the Honolulu Volcanics and Koolau tholeiites nearly define the range of Sr and Nd isotopic ratios in Hawaiian lavas (Fig. 3).
5.1. Honolulu Volcanics
Based on major and trace element abundance data for the Honolulu Volcanics, Clague and Frey  concluded that the spectrum of lavas from alkali olivine basalt to nepheline melilitite formed by varying degrees of melting of a source that was compositionally homogeneous except possibly for Sc, Ti, Zr, Hf, Nb, HREE and Ta contents. The uniformity of Sr, Nd, Hf and Pb isotopic ratios in lavas from the Honolulu Volcanics ([6,7,10,19], this paper) is consistent with this model. Even lavas with unusual compositions (high- and lowTiO2 groups, vent 37 samples with usually low HREE contents, ) are not isotopically distinctive. Consequently, the complex abundance variations of Sc, Ti, Zr, Hf, Nb, HREE and Ta in the lavas of the Honolulu Volcanics probably reflect variations in residual mineralogy rather than trace element abundance variations in the source. Clague and Frey  also concluded that the source of the Honolulu Volcanics was enriched in highly incompatible elements compared to esti-
Fig. 5. Nd evolution diagram. Bulk earth curve (BE) was calculated by assuming that the present-day 143Nd/VUNd of the bulk earth is approximately 0.51264  and 147Sm/l~Nd =0.1967 . Curve labelled "HV" indicates growth of 143Nd/144Nd in the source of the Honolulu Volcanics based on the present-day 143Nd/VUNd = 0.513035 (Table 3) and 147Sm/144Nd ratio of the source estimated by Clague and Frey . Curve HV' indicates growth line for a source with 147Sm/144Nd 10% higher than that estimated by Clague and Frey . Curve labelled "DMS" is an estimate of the evolution of the most radiogenic MORB source calculated after Zindler et al.  assuming that the most depleted MORB source has 143Nd/144Nd = 0.5133, today, the age of earth = 4.6 × 109 yr and a transport coefficient (K) of -4.6×10 11 yr-l. An assumption of this transport model is that the MORB reservoir has differentiated continuously since the earth's accretion. Hence, the DMS curve tracks along a more radiogenic path than most other models for the evolution of the MORB reservoir (e.g. ). Therefore, the intersection of the HV evolution curve with the DMS curve provides an upper limit to the age of metasomatism of the HV source.
mated primitive mantle compositions (e.g. ). However, the relatively high 143Nd/144Nd and low 87Sr/86Sr ratios of the Honolulu Volcanics require that their source region had a long-term history of incompatible element depletion relative to primitive mantle. These interpretations of trace element and isotopic data can be reconciled by postulating a recent metasomatic event in the source region which increased the relative abundances of incompatible elements. A n upper age limit for the metasomatic event can be estimated from the average t43Nd/144Nd ratio, 0.513035 (Table 3) of the Honolulu Volcanics and the 147Sm/144Nd ratio, 0.137, inferred for their source . On a N d evolu-
tion diagram (Fig. 5) these parameters define a growth curve for the source of the Honolulu Volcanics. The intersection of this growth curve with an estimated growth curve for depleted mantle (i.e., high S m / N d , see Fig. 5 caption) provides an upper age limit for the metasomatic event because at greater ages the source of the Honolulu Volcanics would be more radiogenic than any k n o w n mantle reservoir. The age of this intersection is approximately 400 m.y. (Fig. 5). This age is relatively insensitive to inaccuracy in the estimate of the source S m / N d ratio because the high 143Nd/144Nd ratios of the lavas places their source growth line close to that of " m o s t depleted" mantle at t i m e = 0. F o r example, if a 147Sm/144Nd ratio of 0.150, which is 10% higher than that inferred by Clague and Frey , is used to define the growth curve of the source, then an upper age limit of 450 m.y. is inferred for the metasomatism. We emphasize that - 400 m.y. is a m a x i m u m age for the postulated metasomatic event, and that the isotopic data are consistent with metasomatism occurring immediately prior to eruption. These data illustrate the enigma of the metasomatic source hypothesis for alkali basalts. Trace element models require a source enriched in highly incompatible elements with a S m / N d ratio less than the chondritic ratio [8,38,39], but Sr and N d isotopic ratios of alkali basalts typically provide evidence for a source depleted in incompatible elements [9,38-41]. The enriched source compositions inferred from trace element studies can only be reconciled with the isotopic data if the enrichment process (metasomatism) was recent. 5.2. Koolau tholeiites It is established that when normalized to a uniform M g O content, tholeiites from each Hawaiian shield have distinctive geochemical characteristics (e.g. [30-35, 42]). For example, there are significant differences in total iron ( F e O * ) contents between Hawaiian tholeiitic shields, and Koolau tholeiites have the lowest FeO* contents ([33, table 11]; Fig. 4). L a n g m u i r and H a n s o n  postulated that these differences in FeO* content reflect differences in (a) source composition, (b) degree of melting, or (c) pressure and temperature of melt-
152 ing. Specifically, they suggested that the lower FeO* and incompatible element content of Mauna Loa lavas relative to Kilauea lavas reflects a more depleted source for Mauna Loa lavas. Consequently, the anomalously low FeO* content in Koolau lavas may reflect a very depleted source material with relatively low R b / S r and high S m / N d which would evolve with time to a source with relatively low 87Sr/86Sr and high 143Nd/144Nd. However, Koolau tholeiites have higher 87Sr/86Sr and lower 143Nd/144Nd than other Hawaiian tholeiites (e.g. [5,9,10,43-45]; Fig. 3). Therefore, if the low FeO* contents of Koolau tholeiites reflect a mantle depletion event, this event was relatively recent. Alternatively, FeO* contents are decoupled from incompatible trace element abundances ; evidence in favor of this conclusion is the uniform FeO* content of Koolau tholeiites with varying Sr and Nd isotopic ratios. Among Koolau tholeiites the alkali metals, K and Rb, have the largest range in abundance (factors of 8 and 10, respectively, Tables 1 and 2), but their abundances are not positively correlated with abundances of other incompatible elements. Similar discordance has been interpreted as reflecting residual minerals, such as phlogopite, during partial melting (e.g. ), but we believe that the variable K and Rb contents of Koolau lavas reflect late-stage alteration. The Koolau data do not unambiguously support this conclusion because abundances of K and Rb are not strongly correlated with geochemical indices of alteration, such as H 2 0 content and FezO3/FeO; however, in these samples there is only a poor positive correlation between total H 2 0 or H2 O+ content and Fe/O3/FeO. Our interpretation of alkali metal redistribution during late-stage alteration of Koolau lavas is predicated in part on other studies [43,46,47] which show that alkali metals are strongly leached from Hawaiian lavas exposed in a high rainfall environment similar to that of the Koolau shield. We do not understand the cause of the negative Ce anomaly or the 2- to 3-fold enrichment of Y and REE in 69TAN-2 relative to the other tholeiites (Fig. 2 and Table 2). Among the samples studied this sample has the lowest MgO, and Cr contents, but its MgO content is similar to Koolau
tholeiite, 65PB-2  which has normal REE abundances . Moreover, 69TAN-2 is not similarly enriched in other incompatible elements (Table 2). Consequently, we are confident that the Y and REE enrichment in this sample does not result from extensive fractional crystallization or unusually low degrees of melting. Sample 69TAN-2 contains phenocrysts and microphenocrysts of plagioclase, olivine, clinopyroxene, and orthopyroxene and is not petrographically distinctive relative to other Koolau tholeiites. Also, there is no obvious petrographic evidence for late-stage alteration. Although they are not common, there are other Hawaiian tholeiites with unusually high REE contents (tholeiites from Niihau and West Molokai; D. Clague, unpublished). The isotopic heterogeneity of Koolau tholeiites is clearly established even if sample 69TAN 2 is not considered (Table 4 and Fig. 3). The heterogeneity in 87Sr/86Sr contrasts markedly with the Sr isotopic homogeneity of a tholeiitic suite from the vicinity of Honomanu Bay on Haleakala volcano , and of a suite of Kohala basalts . However, the Koolau tholeiites we studied were obtained from diverse parts of the Koolau shield (Fig. 1) and probably represent a longer eruptive period than the Haleakala and Kohala suites.
5.3. Relationship between Koolau tholeiites and Honolulu Volcanics It is possible that the Honolulu Volcanics and Koolau tholeiites formed from unrelated source materials and that there is no petrogenetic relationship between these lava suites . However, the trend for decreasing 87Sr/86Sr and increasing 143Nd/ln4Nd from tholeiites to younger alkalic lavas is typical of several Hawaiian volcanoes; for example, this trend is followed by tholeiitic and post-erosional basalts from Haleakala , East Molokai [12,13] and Kauai . Moreover, alkalic basalts from the shield capping stage also follow this trend at Haleakala  and Kohala [43,47], although alkalic and tholeiitic shield basalts from Waianae, Oahu  and Mauna Kea  have similar Sr and Nd isotopic ratios. It seems likely that these geochemical trends reflect mantle processes which produce the lavas of Hawaiian
153 volcanoes. Thus, it is desirable to develop m o d e l s for H a w a i i a n v o l c a n i s m which can account for these systematic e v o l u t i o n a r y trends in lava comp o s i t i o n s a n d i s o t o p i c r a t i o s at i n d i v i d u a l volcanoes. Specifically, m o d e l s for the lavas forming the K o o l a u R a n g e m u s t explain the following: (a) The isotopic h o m o g e n e i t y of the H o n o l u l u Volcanics a n d isotopic heterogeneity of the K o o l a u tholeiites. (b) The lower 87Sr/86Sr a n d higher 143Nd//144Nd ratios of the alkalic H o n o l u l u Volcanics relative to the K o o l a u tholeiites. (c) The similarity of Sr a n d N d isotopic ratios in the y o u n g e s t lavas, the H o n o l u l u Volcanics, with these ratios in M O R B . T h e geochemical d a t a p r e c l u d e simple m o d e l s which relate the tholeiites to p o s t - e r o s i o n a l alkalic b a s a l t s by different degrees of melting of a c o m m o n source, b y crystal f r a c t i o n a t i o n from a comm o n p a r e n t a l melt at high pressures or b y mixing of two isotopically distinct melts. There are a wide variety of m o r e c o m p l e x m e c h a n i s m s which can be e v a l u a t e d ranging from i n d e p e n d e n t origins for the tholeiites a n d p o s t - e r o s i o n a l basalts  to m i x i n g between at least three isotopically distinct m a t e r i a l s which are irregularly d i s t r i b u t e d in the u p p e r m a n t l e . Here we evaluate the m o d e l p r o p o s e d for H a l e a k a l a v o l c a n o b y C h e n a n d F r e y (; d e s i g n a t e d C - F m o d e l in the following) because it satisfactorily e x p l a i n e d geochemical variations similar to those in the K o o l a u tholeiites and the H o n o l u l u Volcanics. O u r objective is to evaluate the suitability of this m o d e l for the K o o l a u tholeiites a n d H o n o l u l u Volcanics. In terms of Sr a n d N d isotopic ratios the C - F m o d e l is a c o m p l e x two c o m p o n e n t m o d e l with the c o m p o n e n t s derived from (a) d e p l e t e d m a n t l e isotopically similar to n o r m a l m i d - o c e a n ridge b a s a l t ( M O R B ) , a n d (b) an enriched m a n t l e with higher 878r/86Sr a n d lower 1 4 3 N d / l a a N d than M O R B sources. T h e specific c o m p o n e n t s chosen to exp l a i n the H a l e a k a l a lava c o m p o s i t i o n s were (a) melts derived from a M O R B source b y small degrees of melting a n d (b) melts derived from enriched m a n t l e or the enriched m a n t l e itself. Elem e n t a l a b u n d a n c e s b u t n o t isotopic ratios (Sr a n d N d ) in the c o m p o n e n t s are variable b e c a u s e the degree of m e l t i n g of the d e p l e t e d a n d enriched
m a t e r i a l s are variables in the C - F model. Chen a n d F r e y  p o s t u l a t e d that these c o m p o nents f o r m e d as a relatively hot m a n t l e p l u m e r e a c t e d with cooler wall rocks which were pres u m e d to be geochemically similar to the source of M O R B s . They p r o p o s e d two types of physical mixing: (a) m i x i n g of melts f o r m e d b y very small
0 58 - 0 05 [
(~xo~ 9~ ~ao
I I "k
Lines of Constant Mixing Proportions
I i ,"f'- ', ~ ; ~ I ! i "!~ ~\~;~
I ) i
I ~ 1 ~,
o~ ~ I i"~',~'~4~x'~,~ ~ I I ~ 046
~o~p-...~ I I
'~'---.~ "L.\"&\ x
054 052 I 0,7020
dORB SOURCE I [ I I I 07030
HONOMANU THOCE ITES
~ I 0 7040
87Sr / 86Sr
Fig. 6. La/Ce-87Sr/86 Sr illustrating mixing model of Chen and Frey . End-member mantle compositions are bulk earth (BE) and a MORB source (see Chen and Frey [9, fig. 31 for compositions chosen). Vertical lines emanating from BE and MORB source indicate La/Ce ratios as a function of degree of melting (0.05-3% for MORB source; 1-5% for BE). Solid lines between BE and Melts of MORB source are mixing lines and dashed lines indicate constant mixing proportions. The model which is discussed in the text shows that the La/Ce and 87Sr/86Sr variations in Hawaiian basalts can be explained by mixing of melts formed by low degrees of melting of a MORB source with either a mantle of bulk earth composition or melts derived from a bulk earth composition. Fields for Haleakala lavas, Honomanu tholeiites and the post-erosional Hana alkalic lavas are from Chen and Frey . Fields for Honolulu Volcanics and Koolau tholeiites are from Tables 2-4 of this paper. These fields differ from those in fig. 3b of Chen and Frey  because we have plotted only samples whose La/Ce ratio has been determined at M.I.T. (see Fig. 2 caption). Asterisk indicates model source of the Honolulu Volcanics discussed in the text. Koolau tholeiite sample 69TAN-2 was not plotted because of its anomalous REE content. Although the 4 Koolau tholeiite points are sub-parallel to the mixing lines, neither the Haleakala data or data for the Honolulu Volcanics are parallel to the mixing lines; hence, two-component mixing involving components with constant composition cannot explain the data.
154 and variable degrees of melting of the wallrock (MORB source) with the ascending plume (enriched mantle source) to create veined material which subsequently melted to form Hawaiian basalts, and (b) mixing of melts derived by very small and variable degrees of melting of the wall rock with melts derived from plume material. In the latter case, the mixed melts are Hawaiian basalts. Both mixing models account for the geochemical variations in Haleakala basalts if the mixing event involved components of variable composition derived from a MORB source and a relatively enriched source [9, fig. 3]. The variation of Sr isotopic and L a / C e ratios in the Koolau tholeiites can be explained in the context of the C-F model; specifically, four of the five Koolau samples analyzed for REE and Sr isotopic ratios lie near mixing lines between enriched mantle and melts formed by 0.1-0.2% melting of a MORB source (Fig. 6). We have not considered 69TAN-2 because of its negative Ce anomaly (Fig. 2). Because of their relatively high STSr/86Sr and low 143Nd/144Nd ratios ( and Fig. 3) the Koolau tholeiites contain more of the enriched mantle component than other Hawaiian basaltic samples; thus, in the C-F model they constrain the lower and upper limits of 87Sr/86Sr and 143Nd/144Nd ratios, respectively, in the enriched mantle component. The C-F model for Haleakala lavas successfully explained the geochemical characteristics of tholeiitic basalts and the post-erosional alkalic basalts. However, the Koolau tholeiites and the Honolulu Volcanics have additional features which must be explained by a successful model; that is, (a) the isotopic homogeneity of the Honolulu Volcanics which are compositionally diverse, and (b) variations of Pb isotopic ratios in Koolau tholeiites and the Honolulu Volcanics which may require more than two isotopically distinct components [10, fig. 4a]. In the following we address these complications within the context of the C-F model. Based on the 87Sr/S6Sr of the Honolulu Volcanics, we assume that their source had 878r/86Sr ~ 0.70334 (Table 3). We can also estimate the L a / C e ratio in the source because during partial melting the ratio (La/Ce)melt/(La/Ce)re.~idue exceeds unity for all commonly postulated mantle
mineral assemblages; therefore, the smallest L a / C e ratio, 0.46, in the Honolulu Volcanics provides an upper limit for the L a / C e ratio in the source. The sensitivity of the melt L a / C e ratio to degree of melting is indicated by the vertical lines emanating from the MORB and enriched end-members in Fig. 6. Significant L a / C e variations result from < 5% melting, but larger degrees of melting have only small effects on L a / C e ratios; for example, 5% melting of the enriched source ( L a / C e = 0.374) yields a melt with only slightly higher L a / C e ratio (0.388). Consequently, we assume a L a / C e ratio of 0.44 for the source of the Honolulu Volcanics. Using these assumptions and the end-members used in the C-F model, we find that the source of the Honolulu Volcanics could have been created by mixing 97.5% relatively enriched mantle with 2.5% of a melt formed by 0.26% melting of a MORB source (Fig. 6). With respect to our earlier discussion, this mixing event is geochemically equivalent to the recent metasomatic event postulated for the source of the Honolulu Volcanics. The uniformity of isotopic ratios in the Honolulu Volcanics requires that the mixing event created an isotopically homogeneous source which was melted to various degrees to create the compositionally heterogeneous Honolulu Volcanics as outlined by Clague and Frey . Fig. 6 was constructed using the MORB source and enriched mantle compositions suitable for Haleakala lavas . Using the end-member compositions of Chen and Frey  and mixing proportions deduced from the position of the source in Fig. 6, we calculated the incompatible element abundances in the source of the Honolulu Volcanics. These calculated source abundances are remarkably similar to those estimated by Clague and Frey  for the source of the Honolulu Volcanics (Table 5), and they provide additional support for the conclusion that the source of the Honolulu Volcanics was enriched in incompatible elements relative to a primitive mantle composition. The sensitivity of this calculation to the estimate of L a / C e in the source is indicated in Table 5 by results for ( L a / C e ) source = 0.44 and 0.47. We emphasize that Clague and Frey used only trace element abundances to infer source characteristics whereas we used specific L a / C e
155 TABLE 5 Estimates of incompatible element abundances (ppm) in the source of the Honolulu Volcanics Based on trace element abundance systematics in the lavas (from Clague and Frey [8, table 5] for F = 0.11) La Ce Nd Rb Sr Ba Nb Ta Th
2.67 5.44 3.14 60 -96 36 3.1-4.4 0.22-0.33 0.22 Observed mean (from Table 3)
Based on source L a / C e = 0.47 and 8VSr/S6Sr = 0.70334 and mixing model of Chen and Frey  a
Based on source L a / C e = 0.44 and SVSr/S6Sr = 0.70334 and mixing model of Chen and Frey  a
2.41 5.15 2.7 1.6 55 25 3.6 0.19 0.26
2.23 5.05 2.7 1.4 55 22 3.0 0.16 0.22
Based on model
Based on model
End-member compositions and partition coefficient taken from Chen and Frey .
and SVSr/~6Sr ratios coupled with the end-member compositions suitable for Haleakala basalts ( and Fig. 6). There is no a priori reason why these different approaches to estimating source characteristics should yield similar results. Thus, we conclude that the C-F model can explain the incompatible element abundances and isotopic ratios of Sr and Nd in the Honolulu Volcanics and Koolau tholeiites. Although we propose that the source composition of the Honolulu Volcanics was created by mixing of two isotopically distinct materials, the range in Sr, Nd and Pb isotopic ratios established by data from several Hawaiian volcanoes requires more than two isotopically distinct materials [10,45]. Consequently, the C-F model may be adequate for explaining variations in incompatible element abundances and isotopic ratios of Sr and Nd at individual volcanoes, such as Haleakala and Koolau, but it is inadequate for explaining Sr, Nd and Pb isotopic variations on the larger scale of several Hawaiian volcanoes. Possibly, the requirement for more than two isotopically distinct components on a large-scale reflects isotopic variations in the postulated end-members, i.e., MORB sources
and relatively enriched mantle. However, even on the scale of a single volcano variations in Sr, Nd and Pb isotopic ratios require more than two isotopically distinct materials (Loihi, ; Koolau, [10, fig. 4]; Haleakala, C.Y. Chen and S.R. Hart, unpublished). Perhaps Pb isotopic ratios are more sensitive than Sr and Nd isotopic ratios to the presence of other components. An important unsolved problem that we have not addressed is the cause of the renewed volcanism which forms the post-erosional stage after a period of volcanic quiescence. Clague et al.  showed that the quiescent period between the shield building and post-erosional stages was longer for the older volcanoes Kauai and Niihau than the younger volcanoes West Maui and Haleakala. In fact, there appears to be a systematic decrease in the quiescent period from the oldest volcano studied, Niihau ( - 2 . 5 m.y.) to Haleakala ( < 0.4 m.y.), the youngest volcano with a post-erosional series [13, fig. 4]. At Koolau the quiescent period was - 1.2 m.y. Clague et al.  concluded that current hypotheses for the cause of post-erosional volcanism are not plausible.
156 6. Conclusions Compared to other Hawaiian tholeiitic suites, Koolau tholeiites have low iron contents (Fig. 4), unusual Pb isotopic ratios  high 87Sr/86Srratios and low ~43Nd/144Nd ratios (Fig. 3). In addition, Koolau tholeiites are isotopically heterogeneous. Because of the small number of samples studied we cannot evaluate correlations between geochemical characteristics of the tholeiites and their eruption age. However, there are well-defined correlations among geochemical parameters; e.g., inverse trends between 87Sr/86Sr and 143Nd/144Nd (Fig. 3) and between 87Sr/S6Sr and La/Ce (Fig. 6). These correlations can be explained by mixing mantle with bulk earth values of 87Sr/86Sr, 143Nd/144Nd and L a / C e with small amounts of melts derived by 0.1-0.2% melting of a MORB source (Fig. 6). Physically, this mixing event is interpreted as interaction of an ascending, partially molten, geochemically undepleted diapir with its lithosphere wall-rocks which are presumed to be MORB-related . The alkalic lavas of the post-erosional Honolulu Volcanics are remarkably similar in Sr, Nd, Pb and Hf isotopic ratios ([6,7,10,19], this paper), and they, as well as post-erosional lavas from Haleakala , East Molokai [12,13] and Kauai , are isotopically similar to MORB. The isotopic homogeneity of the Honolulu Volcanics coupled with their variable enrichment in highly incompatible trace elements can be modelled by variable degrees of melting of a compositionally homogeneous, enriched mantle (, and this paper). The Sr and Nd isotopic ratios and incompatible element contents of this source are consistent with formation by mixing of components derived from the same end-members that satisfactorily explain Sr and Nd isotopic ratios and incompatible element contents in Haleakala basalts  and Koolau tholeiites (Fig. 6). The geochemical effects of this mixing event are akin to those of the mantle metasomatic processes that are commonly proposed to have affected the source regions of alkalic basalt (e.g. [8,40]). The Nd isotopic ratios in the Honolulu Volcanics require that this mixing event occurred within the last 400 m.y. The Sr and Nd isotopic results for Koolau tholeiites and the alkalic Honolulu Volcanics con-
firm other recent evidence [9,11-13] that alkalic, post-erosional lavas have lower 87Sr/86Sr and higher ta3Nd/144Nd than spatially associated tholeiites. Consequently, an understanding of the evolution of Hawaiian volcanoes requires an explanation for these systematic geochemical variations with age. Models relating Hawaiian tholeiitic and post-erosional alkalic basalt by different degrees of melting of a homogeneous source or different degrees of fractionation from a common parental magma are precluded by such isotopic data. We find that mixing models involving components from MORB sources and enriched sources (i.e., compositions similar to bulk earth estimates) can explain most of the geochemical variations found at individual Hawaiian volcanoes. However, more than two isotopically distinct materials are required to explain isotopic variations in all Hawaiian lavas [10,45].
Acknowledgements Many of our samples were collected by the late E.D. Jackson whose enthusiasm for research on Hawaiian rocks stimulated our research. We thank Dr. S.R. Hart for use of mass spectrometry facilities at M.I.T. and Dr. J.M. Rhodes for use of X-ray fluorescence facilities at the University of Massachusetts and Dr. P. Ila for technical assistance in neutron activation. Nuclear irradiations were made at the M.I.T. Nuclear reactor. This paper was significantly improved by critical comments on earlier drafts by D. Johnston, M. Garcia, W. Leeman and M. Tatsumoto, and was written while Roden was a post-doctoral associate at the Department of Geology and Geophysics, University of Minnesota. This research was supported by NSF grants EAR-7823423 and 8218982.
Appendix 1. Sample descriptions and locations-Koolau tholeiites 69MAK-2 69TAN-2
aphyric basalt, Makawao stream below Palikea, El. 600 ft. plagioclase-olivine-hypersthene clinopyroxenephyric basalt boulder, Moleka stream, northwestof Puu Kakea, El. 1040 ft.
157 sparsely olivine-hyperstheneplagioclase phyric basalt block in tuff, trail on southwest side of Ulupau Head, El. 580 ft. 68KAV-6 chilled border of sparsely olivine phyric basalt dike with secondary quartz veins, southeast side of Puu Pahu, north of Keaahala stream in Kaneohe, El. 60 ft. 66KAPAA-16B chilled border of aphyric basalt dike in Kapaa quarry about 2 miles west of Kailua, El. 400 ft. 68HUL-2 basalt dike, Nuuanu Pali, north of post-erosional Pali vents on old road, El. 920 ft. WW9948 vesicular olivine-hypersthene phyric basalt flow, Makapuu highway saddle. WW9980 plagioclase-olivine-hypersthene phyric basalt flow, west of Waimea Canyon. WW9991 sparsely plagioclase phyric basalt flow, Waiakeakua Valley, El. 580 ft. diabase dike, Palolo quarry. WW10398 WW10403 non-porphyritic basaltic aa flow, Makapuu Point summit. WWl1320 plagioclase porphyritic basalt flow, Moanalua Valley. 70ULUP-8
Note: Wentworth and Winchell  give more detailed locations and petrography of all samples denoted W W x x x x x .
References 1 J.G. Moore, D.A. Clague and W.R. Normark, Diverse basalt types from Loihi seamount, Hawaii, Geology 10, 88, 1982. 2 J.L. Powell and S.E. DeLong, Isotopic composition of strontium in volcanic rocks from Oahu, Science 153, 1239, 1966. 3 E.D. Jackson and T.L. Wright, Xenoliths in the Honolulu Volcanic Series, J. Petrol. 11,405, 1970. 4 D.J. DePaolo and G.J. Wasserburg, Inferences about magma sources and mantle structure from variations of 143Nd/144Nd, Geophys. Res. Lett. 3, 743, 1976. 5 R.K. O'Nions, P.J. Hamilton and N.M. Evensen, Variations in 143Nd/VUNd and 87Sr/86Sr ratios in oceanic basalts, Earth Planet. Sci, Lett. 34, 13, 1977. 6 M.A. Lanphere and G.B. Dalrymple, Age and strontium isotopic composition of the Honolulu Volcanic Series, Oahu, Hawaii, Am. J. Sci. 280-A, 736, 1980. 7 M.F. Roden, F.A. Frey and S.R. Hart, The mantle source for the Honolulu Volcanic Series: Nd isotope evidence, EOS 62, 423, 1981. 8 D.A. Clague and F.A. Frey, Petrology and trace element geochemistry of the Honolulu Volcanic Series, J. Petrol. 23, 447, 1982. 9 C.-Y. Chen and F.A. Frey, Origin of Hawaiian tholeiite and alkalic basalt, Nature 302, 785, 1983.
10 P. Stille, D.M. Unruh and M. Tatsumoto, Pb, St, Nd and Hf isotopic evidence of multiple sources for Oahu, Hawaii basalts, Nature 304, 25, 1983. 11 M.D. Feigenson, A model for the generation of alkali basalts from Hawaii, EOS 64, 902, 1983. 12 D.A. Clague, F.A. Frey and M.H. Beeson, Rare earth element and Sr isotopic evidence for the origin of the East Molokai Volcanics, Hawaii, EOS 64, 902, 1983. 13 D.A. Clague, C. Dao-Gong, R. Murnane, M.H. Beeson, M.A. Lanphere, G.B. Dalrymple, W. Friesen and R.T. Holcomb, Age and petrology of the Kalaupapa basalt, Molokai, Hawaii, Pacific Sci. 36, 411, 1982. 14 K.E. Bargar and E.D. Jackson, Calculated volumes of individual shield volcanoes along the Hawaiian-Emperor chain, J. Res. U.S. Geol. Surv. 2, 545, 1974. 15 R.R. Doell and G.B. Dalrymple, Potassium-argon ages and paleomagnetism of the Waianae and Koolau Volcanic Series, Oahu, Hawaii, Geol. Soc. Am. Bull. 84, 1217, 1973. 16 H.T. Stearns and K.N. Vaksvik, Geology and ground-water resources of the island of Oahu, Hawaii, Hawaii Div. Hydrogr. 1, 1, 1935. 17 H. Winchell, Honolulu series, Oahu, Hawaii, Geol. Soc. Am. Bull. 58, 1, 1947. 18 C.K. Wentworth and H. Winchell, Koolau basalt series, Oahu, Hawaii, Geol. Soc. Am. Bull. 58, 49, 1947. 19 S.S. Sun, Lead isotopic study of young volcanic rocks from mid-ocean ridges, ocean islands, and island arcs, Philos. Trans. R. Soc. London, Ser. A, 297, 275, 1980. 20 J.F.G. Wilkinson, and A.J. Stolz, Low-pressure fractionation of strongly undersaturated alkaline ultrabasic magma: the olivine-melilite-nephelinite at Moiliili, Oahu, Hawaii, Contrib. Mineral. Petrol. 83, 363, 1983. 21 J. Budahn, Geochemical study of Hawaiian basalts, 151 pp., M.S. Thesis, Oregon State University, 1979. 22 F.A. Frey, The origin of pyroxenites and garnet pyroxenites from Salt Lake Crater, Oahu, Hawaii: trace element evidence, Am. J. Sci. 280-A, 427, 1980. 23 S.R. Hart and C. Brooks, The geochemistry and evolution of the early Precambrian mantle, Contrib. Mineral. Petrol. 61, 109, 1977. 24 J.M. Rhodes, Homogeneity of lava flows: chemical data for historic Mauna Loan eruptions, Proc. 13th Lunar Planet. Sci. Conf., Part 2, Geophys. Res. 88 (1983) A864. 25 P. Ila and F.A. Frey, Utilization of neutron activation analysis in the study of geologic materials, Atomkernenergie Kerntectnik, in press. 26 F.A. Frey and D.A. Clague, Geochemistry of diverse basalt types from Loihi seamount, Hawaii: petrogenetic implications, Earth Planet. Sci. Lett. 66, 337, 1983. 27 P. Richard, N. Shimizu and C.J. All~gre, 143Nd/146Nd, a natural tracer: an application to natural basalts, Earth Planet. Sci. Lett. 31,269, 1976. 28 A. Zindler, S.R. Hart, F.A. Frey and S.P. Jakobsson, Nd and Sr isotope ratios and rare earth element abundances in Reykjanes Peninsula basalts: evidence for mantle heterogeneity beneath Iceland, Earth Planet. Sci. Lett. 45, 249, 1979.
158 29 T.L. Wright, Chemistry of Kilauea and Mauna Loa lava in space and time, U.S. Geol. Surv. Prof. Pap. 735, 40 pp., 1971. 30 W.P. Leeman, A.V. Murali, M.-S. Ma and R.A. Schmitt, Mineral constitution of mantle source regions for Hawaiian basalts--rare earth element evidence for mantle heterogeneity, in: Magma Genesis, H.J.B. Dick, ed., Oreg. Dep. Geol. Mineral. Ind. Bull. 96, 169, 1977. 31 Basaltic Volcanism Study Project, Basaltic Volcanism on the Terrestrial Planets, 1286 pp., Pergamon Press, New York, N.Y., 1981. 32 M. Tatsumoto, Isotopic composition of lead in oceanic basalt and its implication to mantle evolution, Earth Planet. Sci. Lett. 38, 63, 1978. 33 R.J. Kirkpatrick, D.A. Clague and W. Freisen, Petrology and geochemistry of volcanic rocks, DSDP Leg 55, Emperor Seamount Chain, in: Initial Reports of the Deep Sea Drilling Project 55, 509, 1980. 34 W.P. Leeman, J.R. Budahn, D.C. Gerlach, D.R. Smith and B.N. Powell, Origin of Hawaiian tholeiites: trace element constraints, Am. J. Sci. 280-A, 794, 1980. 35 D.A. Clague and F.A. Frey, Trace element geochemistry of tholeiitic basalts from site 433C, Suiko Seamount, in: Initial Reports of the Deep Sea Drilling Project 55, 559, 1980. 36 A. Zindler, E. Jagoutz and S. Goldstein, Nd, Sr and Pb isotopic systematics in a three-component mantle: a new perspective, Nature 298, 519, 1982. 37 S.-S. Sun, Chemical composition and origin of the earth's primitive mantle, Geochim. Cosmochim. Acta 46, 179, 1982. 38 F.A. Frey, D.H. Green and S. Roy, Integrated models of basalt petrogenesis: a study of quartz tholeiites to olivine melilitites from southeastern Australia utilizing geochemical and experimental petrological data, J. Petrol. 19, 463, 1978. 39 S.S. Sun and G.N. Hanson, Origin of Ross Islands basanitoids and limitations upon the heterogeneity of mantle sources for alkali basalts and nephelinites, Contrib. Mineral. Petrol. 52, 77, 1975. 40 M.A. Menzies and V.R. Murthy, Nd and Sr isotope geochemistry of hydrous mantle nodules and their host alkali basalts: implications for local heterogeneities in metasomatically veined mantle, Earth Planet. Sci. Lett. 46, 323, 1980. 41 C.J. All6gre, B. Dupr6, B. Lambret and P. Richard, The subcontinental versus suboceanic debate, I. leadneodymium-strontium isotopes in primary alkali basalts from a shield area: the Ahaggar volcanic suite, Earth Planet. Sci. Lett. 52, 85, 1981. 42 C.H. Langmuir and G.N. Hanson, An evaluation of major element heterogeneity in the mantle source of basalts, Philos. Trans. R. Soc. London, Ser. A, 297, 383, 1980.
43 M.D. Feigenson, A.W. Hofmann and F.J. Spera, Case studies on the origin of basalt, II. The transition from tholeiitic to alkalic volcanism on Kohala volcano, Hawaii, Contrib. Mineral. Petrol. 84, 390, 1984. 44 W.M. White and A.W. Hofmann, Sr and Nd isotope geochemistry of oceanic basalts and mantle evolution, Nature 296, 821, 1982. 45 H. Staudigel, S.R. Hart, A. Zindler, C.Y. Chen and D. Clague, Pb; Nd and Sr isotopes of Loihi Seamount, Hawaii, Earth Planet. Sci. Lett. 68, 1984. 46 C.Y. Chen and F.A. Frey, Trace element and isotope geochemistry of lavas from Haleakala volcano, East Maui, Hawaii: implication for the origin of Hawaiian basalts, submitted to J. Geophys. Res. 47 M. Lanphere, 87Sr/S6Sr ratios for basalt from Loihi seamount, Hawaii, Earth Planet. Sci. Lett. 66, 380, 1983. 48 F.A. Frey, A. Kennedy, M.O. Garcia, H.B. West, W.S. Wise and S-T Kwon, Mauna Kea volcano: lava compositions and their evolution, EOS, 1984. 49 J.G. Schilling and J.W. Winchester, Rare earth contribution to the origin of Hawaiian lavas, Contrib. Mineral Petrol. 23, 27, 1969. 50 S.-S. Sun, R.W. Nesbitt and A.Y. Sharaskin, Geochemical characteristics of mid-ocean ridge basalts, Earth Planet. Sci. Lett. 44, 119, 1979. 51 N.M. Evensen, P.J. Hamilton and R.K. O'Nions, Rare-earth abundances in chondritic meteorites, Geochim. Cosmochim. Acta 42, 1199, 1978. 52 A. Zindler, Nd and Sr isotopic studies of komatiites and related rocks, in: Komatiitesl N. Arndt and E.G. Nisbet, eds., pp. 399-420, Allen and Unwin, Boston, Mass. 53 G.J. Wasserburg, S.B. Jacobsen, D.J. DePaolo, M.T. McCulloch and T. Wen, Precise determination of S m / N d ratios, Sm and Nd isotopic abundances in standard solutions, Geochim. Cosmochim. Acta 45, 2311, 1981. 54 S.B. Jacobsen and G.J. Wasserburg, Sm-Nd evolution of chondrites, Earth Planet. Sci. Lett. 50, 139, 1980. 55 A. Zindler, S.R. Hart and C. Brooks, The Shabogamo intrusive suite, Labrador: Sr and Nd isotopic evidence for contaminated mafic magmas in the Proterozoic, Earth Planet. Sci. Lett. 54, 217, 1981. 56 C.Y. Chen, Geochemical and petrologic systematics in lavas from Haleakala Volcano, East Maui: implications for the evolution of Hawaiian mantle, 344 pp., Ph.D. Thesis, Massachusetts Institute of Technology, Cambridge, Mass., 1982. 57 F.A. Frey, M.O. Garcia and W. Wise, Manua Kea volcano: Geochemical characteristics of the transition from tholeiitic to alkalic basalt, EOS 63, 1138, 1982.