We describe a spectral technique to measure the apparent attenuation o
f compressional waveforms recorded during an active seismic tomography
experiment centered at 9 degrees 32'N on the East Pacific Rise. Over
3500 estimates of t are obtained from 0.4- to 0.7-s-long windows alig
ned with crustal P phases, including diffractions above and below a mi
dcrustal magma chamber and Moho reflections which cross the rise axis
at a range of lower crustal depths. We apply a smoothest model inversi
on algorithm to the t measurements to derive images of apparent crust
al Q(-1) both 20 km off axis and within a 16 x 16 km area centered on
the rise crest. The models resolve regions of high attenuation in the
uppermost crust and in a low-Q zone which extends from midcrustal to l
ower-crustal depths beneath the rise axis. Off axis, Q values in the u
pper 1 km average 35-50, while at depths greater than 2-3 km Q is at l
east 500-1000. The high levels of attenuation in the uppermost crust p
robably result from the combined effect of frictional, fluid flow, and
scattering mechanisms. Within 1-3 km of the rise axis, Q increase mar
kedly in the uppermost 1 km to about 65. If the increase in attenuatio
n off axis is entirely due to the similar to 300-m increase in the thi
ckness of layer 2A extrusives required by seismic velocity measurement
s, then Q in layer 2A must be about 10-20. No measurements of Q are ob
tained in the immediate vicinity of the 1.6-km-deep axial magma lens b
ecause no wave paths cross the rise axis through this region. The diff
ractions beneath the magma chamber and the Moho reflections require a
low-Q region, with minimum Q values of 20-50, which extends from no mo
re than 2.5 km depth to the base of the crust. These values are simila
r to laboratory measurements of Q obtained at solidus temperatures and
constrain the low-e region to contain no more than a few percent melt
. The axial magma chamber, which comprises a melt lens and an underlyi
ng crystal mush zone, must be confined to a narrow, I-lan-thick region
through which no rise-crossing paths pass. Inversions for along-axis
structure in the low-Q anomaly show a 20-25% increase in attenuation a
t 2-3 km depth north of 9 degrees 34'N but resolve no such trend at 4-
6 km depth. The along-axis variations may reflect the recent history o
f volcanic eruptions and hydrothermal cooling and do not require syste
matic along-axis variations in magma supply.