Dw. Graham et al., CORRELATED HE AND SR ISOTOPE RATIOS IN SOUTH-ATLANTIC NEAR-RIDGE SEAMOUNTS AND IMPLICATIONS FOR MANTLE DYNAMICS, Earth and planetary science letters, 144(3-4), 1996, pp. 491-503
He-4/He-3 and Sr-87/Sr-86 ratios are highly anti-correlated for a suit
e of seamount glasses from both sides of the Mid-Atlantic Ridge at 26
degrees S; the linear correlation coefficient (r(2)) is 0.99 for 5 loc
alities at 3 different seamounts. The seamounts are located on crust u
p to 2.5 myr old, and have He-4/He-3 as low as 65,400 (He-3/He-4 = 11
R(A)) and Sr-87/Sr-86 as high as 0.70350. These isotopic values are si
gnificantly lower and higher, respectively, than those for basaltic gl
asses recovered from 13 localities along the adjacent ridge axis, wher
e the lowest He-4/He-3 ratio is 92,000 (He-3/He-4 = 7.8 R(A)) and the
highest (87)/(86) Sr is 0.70258. Geophysical studies and the small (1-
2%) degree of helium isotope disequilibrium between vesicles and glass
for three seamount lavas suggest that the seamounts formed on or near
the ridge axis. Because no off-ridge hotspots are present in this are
a, formation of the seamounts probably involved capture by the ridge o
f a passive mantle heterogeneity of 'blob' during rift propagation and
tectonic evolution of the Moore fracture zone. The He-Sr-Nd-Pb isotop
ic results for the seamounts show a general trend toward compositions
observed for the Reunion hotspot in the Indian Ocean. Collectively, th
e seamount and ridge axis results are somewhat enigmatic. In addition
to the highly correlated He and Sr isotopes at the seamounts, a fair c
orrelation exists between He and Nd isotopes (r(2) = 0.70). in contras
t, a correlation between He and Pb isotopes is absent for the seamount
glasses, while an independent, positive correlation exists between He
-4/He-3 and Pb-206/Pb-204 for axial lavas. Apparently, different proce
sses are responsible for the seamount He-Sr-Nd isotope relationships a
nd for the nearby ridge He-Pb isotope relationship. If these relations
are only of local significance and result from complications inherent
in multi-stage mixing of more than two mantle components, then they i
mply that the upper mantle may contain domains with variable 4He/ 3He
ratios, in some cases significantly lower than 80,000 (He-3/He-4 > 9 R
(A)), On the other hand, binary mixing adequately explains the linear
He-Sr isotope trend in the seamount lavas. This linear trend suggests
similar He-3/Sr-86 ratios in the local MORE mantle source and in the s
ource region of the low He-4/He-3 blob, which is most likely the lower
mantle or the transition zone region. This similarity in He-3/Sr-86 i
s inconsistent with a lower mantle 3 He/Sr-86 ratio that exceeds the u
pper mantle ratio by at least a factor of 501 deduced from geochemical
models of mantle evolution. Consequently, rare gas models involving a
steady-state upper mantle and quasi-closed lower mantle may be inappr
opriate if applied at length scales on the order of similar to 100 km,
characteristic of mid-ocean ridge segments.