Magnetic and tectonic studies of the dueling propagating spreading centers at 20 40′ S on the East Pacific Rise: Evidence for crustal rotations

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 98, NO. B8, PAGES 13,835-13,850, AUGUST 10, 1993

Magnetic and Tectonic Studies of the Dueling Propagating Spreading Centers at 20ø40'S on the East Pacific Rise' Evidence for Crustal Rotations

LAURA JEAN PERRAM, 1 MARIE-HELENE CORMIER, AND KEN C. MACDONALD

Geological Sciences and Marine Sciences Institute, University of California, Santa Barbara

We present the results of a magnetic study of a 225 km by 240 km area centered on the dueling propagating spreading centers located at 20ø40'S on the East Pacific Rise. A majority of the data used were collected during a cruise aboard the R/V Moana Wave during which continuous SeaMARC II coverage was obtained. These data were combined with additional data to produce an anomaly map which extends to anomaly-2-aged crust. A three-dimensional inversion in the presence of bathymetry was carried out for the area. The resulting magnetization distribution was interpreted and compared to side scan sonar and bathymetry data sets in order to deternfine the recent history of the discontinuity. The results indicate consistent asymmetric spreading faster to the east, discontinuous high magnetizations in the discordant zone associated with the discontinuity, and southward migration of the feature at a rate of 90-100 mm/yr between Jaramillo and Brunhes time (0.95 to 0.73 Ma) with slowing during the Brunhes to less than 10 mm/yr. There is also an overlapping Jaramillo isochron on the west flank and a gap in that isochron on the east flank indicating a transfer of crust during this time period from the Nazca to the Pacific plate. In addition, areas of oblique lineations possibly representing rotated crust were modelled using an inverse method which enables the specification of a nonuniform magnetization unit vector. Results from this second model support the presence of highly rotated pre-Brunhes Nazca crust within Brunhes Pacific crust which has been deformed by bookshelf faulting. This indicates at least two episodes of crustal transfer from the Nazca plate to the Pacific plate. The discontinuity appears to mark the boundary between rigid plate tectonics to the north and deformation within the Nazca plate between the discontinuity and the Easter microplate to the south. The detailed history of the discontinuity involves dueling propagation with a great deal of variation in the amount of overlap of the two ridges as well as inward and outward cutting and abandonment of the tips of both ridges.

INTRODUCTION

In September and October 1987, a SeaMARC II (SMII) survey (MW8710) was completed of the 15 km offset of the East Pacific Rise (EPR) at 20ø40'S (Figure 1). In addition to SMII side scan sonar and bathymetric data, we collected gravity and surface magnetic data. This paper focuses on the magnetic data and our interpretation of these data. Our objectives were to determine the magnetic character of this discontinuity through time, to constrain its recent history and to correlate our findings with other data sets.

The nontransform nature of this discontinuity was first discussed by Rea [1978]. His survey (Nazca Plate Project data) revealed a broad V-shaped swath of tectonically and magnetically disrupted crust indicating southward migration of the discontinuity which has existed since at least anomaly 2 time (1.77 Ma). Rea found that spreading has been asymmetric in the area (faster to the east) lbr at least the past 4 m.y. and suggested that the discontinuity had developed after a period of differential asymmetric spreading or a ridge jump. Further analyses of the Nazca Plate Project data by Macdonald et al. [1988] and Lonsdale [1989] corroborate this origin for the discontinuity but do not resolve the reason tbr this initial period of differential asymmetric spreading.

Macdonald et al. [ 1988] mapped the axial expression of the discontinuity with high resolution Sea Beam and Deep-Tow systems and found that both ridge segments associated with this

1Now Laura Jean Penvenne, at Triton Technology, Incorporated, Watsonville, California.

Copyright 1993 by the American Geophysical Union.

Paper number 92JB02913. 0148-0227/93/92JB-02913505.00

discontinuity have propagated and retreated in the past 0.1 m.y. with net southward motion. Because of this behavior, the feature

is referred to as a dueling propagator. The amount of overlap has changed rapidly due to abandonment of ridge tips as the ridges cut inside or outside of themselves [Macdonald et al., 1988]. They also note that the 20ø40'S discontinuity is located at a maximum in the axial depth profile which is the deepest point along the 1100 km segment from the Garrett transform to the Easter microplate (Figure 2). In conjunction with the present study, other work [Cormier et al., 1988] suggests that the location of the present discontinuity may be a region of cold mantle or decreased mantle upwelling.

DATA COLLECTION AND PROCESSING

The combined MW8710 and Nazca Plate Project magnetic data sets provide coverage of a 225 by 240 km area centered on the 20 ø40'S discontinuity. Data from the Protea 1, Pascua 2, and Ariadne 2 cruises provide dense coverage along the axis and in several small survey areas (Figure 3). These data were supplemented with additional lines crossing the area and the appropriate International Geomagnetic Reference Field (IGRF) [IAGA Division I Working Group 1, 1987], based on the 1985 model which includes secular variations, was removed from each

data set. Navigational en'ors were minimized by shifting based on swath and conventional bathymetric data. Large crossover errors remained due to an absolute shift (-60 nT) of the MW8710 data relative to the remainder of the data. This shift is

presumably due to inadequacy in the IGRF model. In order to correct this problem we performed a least squares inversion of the crossover residuals [Prince and Forsyth, 1988], determining the magnitude of constants to be added to the values along each line of the MW8710 survey, with a final result of minimizing crossover errors for the entire area. The resulting RMS crossover

13,835

13,836 PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS

tudy Area 162 mm/yr _

-- _1EASTER MICROPLATE

L 110ø I I

151 mm/yr QUEB

YA

PACIFIC -"WLKES PLATE

• GARRE'I-I'

I I

NAZCA

PLATE

Accretionary Plate Boundaries:

Tr•em Propagating riffs verlapping Spreading

rs Faults

100 ø 90 ø 80 ø

20 ø

Fig. 1. Location map showing the approximate boundaries of the study area.

N

o t- o o L•

• 0

0 0 0

' 15o'00 .... 20o00, S

Fig. 2. Axial depth profile from the Garrett transform to the Easter microplate with a maximum at the 20ø40'S dueling propagators.

Ill

¸¸

error is 34 nT and can be attributed primarily to diurnal variation. Following the addition of smooth hand-drawn contours in areas of sparse coverage, the combined data set was gridded at a 1-km interval

The IGRF is poorly constrained over the southeast Pacific. Therefore, since the study area covers a somewhat even distribution of normally and reversely magnetized seafloor, the residual long-wavelength component was removed from the field in the form of a least squares fit plane through the gridded data set. This local correction is given by the relation

-1097.9 + 41.675 * (latitude) - 17.364 * (longitude)

where latitude and longitude are in negative decimal degrees and the correction is in gammas. The final version of the magnetic anomalies is shown in Figure 4.

A grid of bathymetry with a 1-km interval was produced over the same area as the magnetic anomaly grid (Figure 5). The SMII bathymetry with its 10-km swath width was the major source of data for this map. A grid of these data was combined

with grids of Sea Beam data which are part of a data synthesis developed at the University of Rhode Island [Tighe et al., 1987]. Additional off-axis Sea Beam and conventional echo sounder data

were also used to constrain the bathymetry.

MODELING

The magnetic anomaly and bathymetry grids were used as inputs for a three-dimensional Fourier inversion [Parker and Huestis, 1974; Macdonald et al., 1980; Miller and Hey, 1986] removing the effects of bathymetry and skewhess. We assumed a 1-km-thick magnetized layer which follows bathymetry and a uniform magnetization direction. Also inherent in the inversion process is the assumption that the measured anomaly is due entirely to remanent magnetization in the crust. A straightforward mirroring of the bathymetric and magnetic field grids prior to inversion was found to introduce large artifacts in the magnetization sQlution. Indeed, because the magnetic lineations are striking approximately 010 ø, the mirrored field

PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S. EVIDENCE FOR CRUSTAL ROTATIONS 13,837

19o20 '

19%0'

20ø00 '

20o20 '

20040 '

21ø00 '

21ø20'S

114ø40 ' 114ø20 ' 114ø00 ' 113ø40 ' 113ø20 ' 113ø00 ' I 12ø40'W

Fig. 3. Track chart showing the density of magnetic coverage in the study area. The shaded areas represent regions of very dense coverage. With the exception of the MW8710 survey, we have omitted tracks in the very dense regions. Bold lines indicate MW8710 coverage and lighter lines indicate additional coverage of magnetic data. Only portions of lines containing magnetic data used in the compilation are shown. In some areas, bathymetric coverage is more complete than indicated by the magnetic coverage.

introduces angular discontinuities about the axes of symmetry which could only be produced by geologically unreasonable magnetization distributions. This problem has been resolved by selecting grid boundaries which are approximately parallel to the magnetic lineations and which encompass the initial grid. Both the bathymetry and the field anomalies were extended with grid- parallel contours to fill the rotated grid. The magnetization solution was computed iteratively, and band-pass filtering was applied prior to each iteration [Schouten and McCamy, 1972]. Wavelengths shorter than 3 km and longer than 650 km were fully cut, while those greater than 6 km and smaller than 325 km were passed unattenuated.

Because of the inherent nonuniqueness of the inversion process it was necessary to select a geologically reasonable solution fi'om the infinite number of possible solutions to the problem. An annihilator is a magnetization distribution which causes zero field for a given bathymetry distribution. Because it causes zero field, any multiple of the annihilator may be added to an inversion solution. For the present study area we added 2 times the

annihilator so that the midpoint of the peak-to-trough amplitude of the reversal signals con'esponds approximately to the zero contour (Figure 6).

In addition to the conventional inversion, we performed inverse modeling in which we specified a nonuniform magnetization unit vector. The method for this modeling was developed by S. Huestis and S. Miller and is essentially the same as the ordinm'y inverse modeling with the added flexibility of specifying a nonuniform magnetization unit vector. We specified magnetization unit vectors rotated in the horizontal plane within limited areas infen'ed from the side scan and bathymetric data to be rotated. We then inverted with the same parameters as discussed above.

RESULTS

Magnetic Field Anomalies

In the magnetic anomaly map (Figure 4) the Brunhes and Jaramillo anomalies are clear and well defined. A fight step of

13,838 PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS

Fig. 4. Magnetic field anomaly map for the study area. The map was contoured from a 1-km interval grid, and a least squares fit plane was removed. The contour interval is 50 nT. Areas with positive field are shaded. Anomalies are identified as B/M (Bruhnes / Matuyama reversal, 0.73 Ma), J (Jaramillo event, 0.95 Ma) and 2 (1.77 Ma). Heavy lines represent the ridge axis and the 2900-m bathymetric contours which best outline the major seamounts and ridges (compare Figure 5). The location of the strong axial anomaly corresponds closely to the location of the ridge axis as defined from the bathymetry. Note the disturbed field associated with discontinuity area, the overlapping Jaramillo on the west flank, and the associated missing anomaly on east flank. Also note the bull's-eye-shaped negative anomaly near 20ø30'N, 114ø20'W within the west flank Brunhes positive anomaly.

approximately 15 km exists in both anomalies. There is clear overlap in the Jaramillo on the west flank and a gap in the anomaly on the east flank of the rise axis. There are several additional distinctive features in this map. There is a roughly V- shaped zone of distin'bed anomalies associated with the 20ø40'S discontinuity which is manifested within the central anomaly as a broad cross-strike oriented negative anomaly. This anomaly is

located at the southern end of disturbed areas in the side scan

sonar and bathymetry. To the north of this anomaly, on the west flank of the rise axis,

is a bull's-eye-shaped negative anomaly within the central anomaly. This is the most negative anomaly within the study area, and it is located within a region of oblique lineations in both the bathymetry and the side scan sonar.

PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS 13,839

19030'-- ß •11

'!:' ,'"'t• -I \ ,,, ,•

o .... _

, ,.,:.:',,,• •,,•'•1'• •.e,,[• ,:,::•. • " •1 f. a"•[l' :',, o I'"' '"i' 0 '"' * ..:, .,. .... .• ,'•). X,L,,//E •,, ]• o •, ., ... E, 20 e O0 ................ ,,• .................. I't, • ,•li,[•'"'FJ/ • • • • •['"•11'. • .....

, ,: ,,•...• '•,....,, , ,. .: ,, • E,,'*'?rc •,'•i ,, .... ,, ,-

,'""",,, ,• •i, *, ....... '",,,,,•,,;,,,•'•1" •,• * > -*' :,::•"*'" I' •l *"': "'•'• ' ,',•' •*" • •q' ' / ':'• .', ".," i',,,' .... ."' ...... "' ....... .. ."",•' •. '...•'""'.'.'.."• I'"",. 'J '"•[l, [ •'.• ,, .'-Z' ', .... • ,• ,, _ O• • ,• g' *" ,11•';,,, ,,',';,,;? "' ... "' ,,,- •; ,,,,,,,', • ,,:' • II :,• ,•,, ;,:, •o, ..... ,,,: ,,,,:,•, • ...... ,,,, ,,

•llJ "" • [•li •"'• ::'" I •,[[, [[n[,[ll[ ¾,,,,n• .' ,rlq, "" '• "' [., ,('",., ,,....,. , . ;'",,:,•,,,,' ,;,i:[•::• .... • r;n •,• ,, , I'" [W •:,'" .[[.[,:• [ • ' •[.. . j"'[":[n'i. ' ,,3 • " ,,,, • ' '

,." • [[.,,l'i[,.... [' ' '"" .4,....1 I"' •0ø•0 ..

t• ,'• ........... .... d,'d'"" :",,,',, t, ] • TII•

• ..1•:1 ............... •'rll•"ah •1 n n"• I;I •,,.

............ ,,:,,• •,•,• , . •r ...... •,•,, no"

•'oøøl,,,,, ,,i•,•j. •.••,,•"•,,'" ,,•,,,,,,",,,,'•,' [r"'":• "?' '........................[i..t•l]'• ':• ...;"•[,,'[" 0 •, a ,,,,11"gi' , 0 I •,,,:[[:,,[,,' ........ '['[•' ......... '( )0, o •

'• ,, ..... •,, . •,% ß , ' i14o•0 ß 114000 '

I

113 ø 30'

i, ,i

, []

1 13 ø 00'

Fig. 5. Contour map of bathymetry from a 1-km interval grid. Contour interval is 100 m. The smaller box indicates the map area of Figure 8. The larger box indicates the map area of Figure 9. The bold lines denote the EPR axis. Relief shallower than 3000 m is in white, and relief deeper than 3200 m is in the darker shades.

The central anomaly between 20 ø and 21øS is excessively skewed, with very high amplitudes at the western Bruhnes/Matuyama reversal boundary and low amplitudes at the eastern Brunhes/Matuyama reversal boundary. This large skewness is well explained by the presence of the right-stepping kink in the magnetic lineation pattern associated with the ridge axis discontinuity. A forward model of the magnetic field using a simplified magnetization distribution model for the study area clearly reproduced this excess skewness in the discontinuity discordant zone.

The present axis can be identified in the magnetic field anomaly map by the strongly positive anomaly near the center of the Brunhes anomaly [e.g., Klitgord, 1976]. This magnetically

identified axis corresponds nearly exactly to the ridge axis as identified in the bathymetry.

Inferred Magnetization Distribution

Our preferred magnetization distribution solution of the conventional inversion (Figure 6) was obtained from the field distribution shown in Figure 4 with two annihilators added as stated above. Using this solution and the time scale of Hadand et al. [1982], spreading rates were measured 25, 50, and 75 km away from the axial disturbed zone to the north and 25 and 50 km away from the zone to the south (Figure 7). The Bruhnes/Matuyama reversal boundary (0.73 Ma) was taken at

13,840 PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS

12

19',,'30

20•00

20•30 ,

21 ø00

i

, ,.

BIMJ

ß

! 14o30 ' aim aim

JJ

.5 .• .......... I,[ • .•.. ,,, ,,• ,..,,. Ii, •,'•;,,,, I'"" ' '" , .... ',", ,•",,,•,, •m •',•1 ",,,,,,•i•ll,,,•:• ,

II1',, , , , •; ; .

............ •?.•,..'. :• ...... •,,. ,' h,, i"• ';½"'? ..;" . •,,, "'; •:,' ' ' ,, i .... , .,**"•' •I• ," • g• • ,

",•

..... •m ='"' •" '" I IIId: ' h,

• 13 ø 30' Fig. 6. Magnetization distribution obtained by three-dimensional Fourier inversion methods. A 1-km-thick magnetized layer, following bathymetry was assumed, and two times the annihilator was added to the solution. A band-pass filter was applied where wavelengths less than 3 km and greater than 650 km were cut, while wavelengths greater than 6 km and less than 325 km were passed unattenuated. Magnetic anomalies, seamounts and the ridge axis are identified as described in the caption of Figure 4.

midamplitude of the peak-to-trough magnetization reversal, and the mid-Jaramillo event (0.95 Ma) was measured at the peak amplitude. Distances fi'om the axis were measured along 103 ø, the mean spreading dh'ection in the area.

The resulting rates shown in Figure 7 indicate consistent asymmetric spreading with the Nazca plate accreting crust more rapidly throughout the study area. Over the past 1 m.y., spreading to the north of the discontinuity is 3-4% faster than spreading to the south of the discontinuity. With the exception of an apparent increase in spreading for the time period between the

Jaramillo and the Brunhes, there has been a steady decrease in spreading rates since anomaly 2 [Macdonald et al., 1988, Figure •2].

In the solution shown in Figure 6, normal magnetization intensities on the west flank north of the discontinuity are higher by about a factor of 3 than no•xnal magnetizations elsewhere in the study area. We believe that this area of higher magnetization is a real feature of the magnetization distribution which may have been enhanced in the inversion process. Indeed, forward modelling of the magnetic field for a 1-km-thick simplified

PERRAM ET AL.: EAST PAC3FIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS 13,841

Fig. 7. A schematic diagram indicating interval spreading rates 25, 50, and 75 km north and 25 and 50 km south of the discontinuity. Spreading half rates are in millimeters per year. Distances from the ridge axis have been measured in a direction parallel to the NUVEL-1 mean spreading direction (103 ø) ancJ converted to spreading rates using the geomagnetic time scale of Ha.r'l•nd et al. [1982]. Spreading rates are consistently asymmetric, faster to the east.

magnetization model indicates that the main features of the amplitude variations in the observed field could be created by a magnetization variation of about 50% across the ridge north of the disco, ntinuity. Some asymmetry is also observed in magnetization solutions of the Hump area located to tl•e north of the present study area and encompassing the northern portion of the present study area (M.-H. Cormier and K.C. Macdonald, East Pacific Rise 18øS-19øS ß Asymmetric spreading and ridge reorientation by ultra-fast migration of axial discontinuities, submitted to Journal of Geophysical Research, 1992]. Although we will not attempt to interpret this feature, there are severa! other observations and interpretations which can be made regarding the magnetization distributions in this solution.

On the west flank of the rise axis, as in the anomaly map, our magnetization distribution indicates an overlap of the Jaramillo isochron. This overlap is matched on the east flank by a nan'ow or missing Jaramillo isochron. This suggests a transfer of crust from the Nazca to the Pacific plate during or immediately following the Jaramillo event.

Both the edge of the Brunhes age crust and the Jaramillo isochron are offset by essentially the same distance as the present-day offset of 15 km. The present-day ridge axis to the south of the discontinuity is oriented 008 ø , while that to the north of the discontinuity is oriented 013 ø . The orientation of the northern ridge segment is in excellent agreement with the ridge axis orientation predicted by the NUVEL-1 present day plate motion model [DeMets et al., 1990] (Table 1). The southern ridge segment, on the other hand, has an orientation which differs 5 ø in a counterclockwise direction from the orientation predicted by NUVEL-1. This difference of 5 ø is decreased to less than 2 ø going back in time to the Jaramillo (0.95 Ma). At that time the southern segment was oriented 011ø while the northern segment was oriented 013 ø .

In general, magnetization intensities increase toward the

discontinuity. On the axis, the highest intensities are located approximately 20 km behind the tip of the northern ridge segment. This is consistent with the higher resolution work of Macdonald et al. [1988] in this area as well as studies of other migrating discontinuities [Sempdrd et al., 1984; Perram and Macdonald, 1990; Sempdrd, 1991].

In this solution there remains a strong, apparently negative magnetization surrounded by high positive magnetizations on the west flank within Brunhes age crust. These features are located in an area which is recognized in the bathymetry (Figure 5) and side scan sonar (Figure 8, western area) as having oblique lineation orientations. This area consists of a central subarea of

lineations 60o-70 ø oblique to regional trends which is embedded in a subarea where lineations are approximately 32 ø oblique to regional trends. Because of these features, we believe that the imposed condition of a uniform magnetization unit vector is inadequate in this area.

Inferred Magnetization Distribution with Rotations

Due to the presence of oblique lineations in the side scan sonar and the large amplitude magnetizations in the solutions of the conventional inversion, we perfo•xned an inversion with rotated magnetization unit vectors as described above. Figure 8a indicates areas of oblique lineations as identified from the side scan sonar mosaic and bathymetry. There are two major areas' an area on the west flank of the ridge axis corresponding to the location of the bull's eye negative magnetic anomaly described above, and an area near the present axial discontinuity. In Figure 8, the two areas are divided into subareas of consistent lineation

orientation. These subareas were modeled with magnetization unit vectors rotated in the horizontal plane. All rotations were counterclockwise with the magnitude indicated in each subarea of Figure 8a. In the west flank area of oblique lineations, the subarea labeled "20 ø" was not modeled as rotated since lineations, while averaging 200 of obliquity, are inconsistent in their orientation and, therefore, probably do oot represent a coherent block. Figure 9b shows our preferred solution for the area immediately surrounding the discontinuity. Two annihilators were added to the inversion solutiQla so that the magnetization distribution away from the rotated blocks is identical to that of the conventional solution (Figure 6).

In this solution, the magnetization intensities in the west flank area of oblique lineations are slightly lower than in the conventional solution. The west flank negative magnetization is shifted to a position directly over the subarea identified as highly oblique (at least 600 ) in indicates that this crustal block is both negatively magnetized and rotated at least 60 ø. A portion of the area to the south of the highly rotated subarea also seems to be negatively magnetized and has associated with it lineations up to 20 ø oblique to surrounding fabric. As noted above, this zone was not modeled with a rotated magnetization unit vector due to the variable orientation of the lineations. Along a trend which is parallel to relative plate motion these west flank zones of apparent negative magnetization con'espond to a region of east flank Matuyama age crust which is too narrow. This suggests that the west flank blocks were transfen'ed from the Nazca plate. The timing and nature of the dueling propagation which resulted in the transfer of these negatively magnetized subareas will be discussed further in the next section.

In the axial area of oblique lineations, the presence or absence of rotated magnetization unit vectors seems to make little difference in the magnetization distribution solution. This

13,842 PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS

TABLE 1. Relative Plate Motions

Northern Ridge Segment Southern Ridge Segment

Model Ridge Axis Full Spreading Ridge Axis Full Spreading Orientation Rate, mm/yr Orientation Rate, mm/yr

NUVEL-1 013 ø 155 014 ø !55

3P 016 ø 153 016 ø 154

Measured 013 ø 155 008 ø 150

Current Pacific/Nazca plate motion for ridge segments to the north (20øS) and south (2'1øS) of the dueling propagator. Spreading rates and ridge axis orientations are as predicted by the NUVEL-1 global plate motion model of DeMets et al. [1990] and the 3P Pacific-Easter-Nazca model of Naar and Hey [1989]. Measured rates and orientations are from this study.

20o20 '

20030 '

45 ø

20040 '

114ø20 ' 114ø10 ' 114øOO ' 113%O'W

Fig. 8a. Tectonic chart for the central portion of the study area as outlined in Figure 5. Fault and lineation identifications are from the side scan sonar mosaic shown in Figure 8b. Subareas with bold outlines were modeled as rotated in the magnetics (see Figure 9). The numbers indicate the amount of counterclockwise rotation in the horizontal plane that was modeled for each subarea.

20ø50'S

indicates that the oblique lineations here cannot be unequivocally labeled as rotated material. If the oblique lineations do represent rotated material, the magnetic signal associated with this material is not significantly different from that associated with unrotated material in the same area. Although the side scan sonar data suggest rotation in both the west flank and axial areas, this magnetic modeling supports the interpretation of rotation in the west flank area but is inconclusive in the axial area.

HISTORY OF THE DISCONTINUITY

The formation of the discontinuity of the EPR at 20 ø40'S occurred at least 2 m.y. ago during an episode of differential

asymmetric spreading [Rea, 1978; Macdonald et al., 1988]. Magnetic data suggest that the feature has experienced southward migration in the last 2 m.y. [Rea, 1978; Macdonald et al., 1988]. However, in the present study, it is evident that within Brunhes age crust the discordant zone associated with the discontinuity lies along a trend which is essentially pm'allel to the relative plate motion, suggesting that there has been little net migratitn of the discontinuity during this time period. Instead, there has been

/dueling propagation of both of the segments [Macdonald et al., 1988].

The presence of a double Jaramillo isochron on the west flank indicates that crust was removed from the Nazca plate and added to the Pacific plate as the discontinuity migrated southward

PERRAM ET AL.: EAST PACWIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS 13,843

ß

. ...

ß .

Fig. 8b. SeaMARC II side scan sonar mosaic from the central portion of the study area as outlined in Figure 5. Tectonic interpretations are shown in Figure 8a.

during and after the Jaramillo event. The large overlap may be due to very rapid propagation during that period, a large amount of overlap of the ridge axes existing at the time, linking of the segments at the southern end of the overlap, or any combination of these processes. One possible scenario for the development of overlapping Jaramillo age crust is shown in Figure 10. In this scenario, there was a small amount of overlap between the two ridges at the beginning of Jaramillo time. Rapid propagation of both ridges proceeded during the Jaramillo, but by the end of the time period the southern ridge had retreated, relocating the discontinuity at the southern end of the previous overlap. The difference in the location of the offset at the end of Jaramillo time

and at the beginning of Brunhes time suggests rapid migration 0.9-0.7 Ma of 90-100 mm/yr. In the nonunique model shown in Figure 10, much of this migration was accommodated by discrete relocation events for both ridge tips.

The west flank area of rotated crust undoubtedly records at least one episode of propagation and the entrapment and rotation of crust within an overlap region. In the simplest model of propagation with no overlap, magnetic stripes of the transfe•xed crust could be expected to be rotated 60 ø to 65 ø if the propagation rate is approximately equal to the spreading rate [Hey et al., 1986]. This explanation is inadequate for the west flank area due to the presence of varying amounts of rotation. Another possible

history for this area includes the steady southward propagation of the northern segment with a large amount of overlap and episodic inward cutting of the southern dying ridge as described by Wilson [1990]. In this scenario, the west flank rotated area would

represent one packet of crust cut off by the dying ridge which includes highly rotated, transferred crust embedded within less highly rotated crust which was formed within the region of overlap [Wilson, 1990]. However, due to the dueling and overlapping nature of this discontinuity, we believe that th•s area records more than one period of propagation. One possible origin for the zone is shown in Figure 11. Figures 10, 11, and 12 were created using the kinematic modeling program of Wilson [ 1990] with the time scale of Harland et al. [ 1982]. The models are imperfect and nonunique, and we feel that the processes indicated by these m'e more important than the exact details of the history. The history shown in Figure 12 includes rapid propagation of both segments (0.76 Ma), linking of the two ridge segments (0.68 Ma), and renewed southward propagation followed by retreat of the southern segment causing the rotated block to be rafted to the west flank (0.56 Ma). Renewed northward propagation (0.40 Ma) may have fo•xned the small oblique ridge bounding the southeast side of the m'ea of oblique lineations (Figures 5, and 8). Regardless of the sequence of events, it is evident that the most highly rotated block and the

13,844 PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS

20 ø 20'

20 ø 30'--

20 ø 40

Ridge 3'

• " Ridge 2

Ridge 4

20 ø 50 114o20 ' 114 ø 10' 114o00 ' 113ø50 '

Fig. 8c. High-resolution bathymetry for the central portion of the study area as outlined in Figure 5. The contour interval is 50 m and the map was created from a 200-m grid. Areas shallower than 3100 m are in white, areas deeper than 3300 m are in the darker shades. Ridge identifications are after Macdonald et al. [1988]. Tectonic interpretations and the side scan mosaic for this area are shown in Fig. 8a and Fig. 8b respectively.

block to the southeast are negatively magnetized and were originally N azca crust. It is clear that these blocks were transfe•xed to the Pacific plate and rotated during propagation of both segments and that the episode ended in the retreat of the southern segment.

The most recent history (< 0.1 Ma) of this discontinuity is extremely complex and the basic scenario is described by' Macdonald et al. [1988]. Our broader coverage of the area resolved few variations from the model of Macdonald et al.

[1988] in the evolution of the feature over the past 0.1 m.y. Figure 12 shows a history for the past 0.1 m.y. which is slightly altered fi'om the history of Macdonald et al. [1988]. At 0.1 Ma, there was a large amount of overlap and extension at the tip of the southern ridge segment which formed the deep basin to the east of Ridge 1' (Basin A as defined by Macdonald et al. [ 1988]). At this time, the tip of the northern ridge advanced to foden Ridge 4, while the southern ridge cut inside of itself to form Ridge 2 (0.06 Ma). In the most recent history of the discontinuity it is clear from the side scan (Figure 8) and bathymetric (Figm-e 5) data that the rotated area was partially divided as the present northern ridge axis (Ridge 3) cut inside of Ridge 4. This occurred at 0.06 Ma in our model as the tip of the southern ridge advanced cutting to the outside of the previously formed hole.

As Macdonald et al. [1988] discuss in their "note added in proof", the most recent change in the southern segment is the inward cut which abandoned ridge 1' in favor of ridge 1. In our model, this final change occmxed 0.02 Ma.

DISCUSSION

Regional Features

In this work, we have focused on the local features of the

magnetics which have helped us to trace the history and some of the processes occun'ing at the 20ø40'S discontinuity. However, the broader-scale magnetic features of the study area also have important implications for processes which may be occurring along this ultra-fast spreading (155 mmIyr) segment of the East Pacific Rise.

Spreading rates in the study area indicate consistent asymmetric spreading faster to the east and a reduced spreadbig rate south of the discontinuity. Rea [1978] documented asymmetric spreading faster to the east in several areas along the EPR south of the Gan'ett transform. The asymmetry is therefore not directly related to the discontinuity at 20ø40'S but has a more regional origin. One origin for prolonged, broad-scale asymmetric spreading was suggested by Stein et al. [1977]. They

PERRAM ET AL.' EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS 13,845

20 ø ! O' ,

20 • 20'

20 • 30'

20•40 ß

20•50 ß

114'40' 21 eO0'

114e30 ' 114e20 ' i 14elO ' 114e00 ' 113e50 ' 113e40 '

Fig. 9a. Inversion solution without rotations of the magnetization unit vector in the central area outlined in Figure 5. Two annihilators were added to this solution. Heavy lines indicate the outlines of rotated bodies shown in Figure

argue, based on a fluid mechanics model, that the lateral motion of a ridge axis relative to the underlying asthenosphere favors faster spreading of the "trailing" flank. Indeed, the southern EPR is migrating westward, and in the study area the east flank is spreading faster. Davis and Karsten [1986] propose that another effect of rapid lateral ridge motion is the initiation of mantle upwelling in advance of the migrating spreading center, accounting for a greater abundance of seamounts on the "leading" flank of the ridge. A subsidiary effect of an asymmetric mantle upwelling pattern might also be the anomalously low subsidence rate of the Pacific plate ("leading" plate) in the study area. This asymmetry in flank subsidence rates was studied by Cochran [ 1986] who suggests, as a cause, a lateral temperature gradient of 0.05-0.10 ø C/kin in the underlying mantle.

The difference in spreading rates to the north and south of the discontinuity (Figure 7 and Table 1) may be the result of compression in the Nazca plate associated with the northern boundary of the Easter microplate. An anomalous southward decrease in spreading rate between 22øS and the northern boundary of the Easter microplate has been attributed to nom'igid behavior of this portion of the Nazca plate [Searle et al., 1989; Naar and Hey, 1989 & 1991; Rusby and Searle, 1991]. The spreading rates measured here suggest that nonrigid behavior

extends as far north as the 20ø40'S discontinuity. The ridge axis orientation and spreading rates to the north of the discontinuity are virtually identical to those predicted by NUVEL-1, indicating that the assumption of rigid plate tectonics is valid in this area. Therefore, the 20ø40'S discontinuity may mark a transition between rigid plate tectonics to the north and nonrigid deformation of the Nazca plate to the south.

As mentioned previously, the ridge segment to the south of the 20'40'S dueling propagator has undergone approximately 5* of counterclockwise rotation since anomaly 2 time (1.77 Ma). Two right-stepping overlapping spreading centers (OSC) within this segment may have been initiated or grown in offset as a result of this rotation. As with the spreading rate anomalies discussed above, this rotation may be a response to Nazca plate compression north of the clockwise rotating Easter microplate [Searle et al., 1989; Naar and Hey, 1989 & 1991; Rusby and Searle, 1991]. Assuming perpendicular spreading, north-south compression of the Nazca plate should cause a counterclockwise rotation of the least compressive stress and therefore, the ridge axis.

SeaMARC II side scan data collected in 1987 (MW8710) along the axis of this segment of the EPR indicate a progressive counterclockwise rotation of the ridge axis from the 20ø40'S

13,846 PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS

20e!O ß

114040 '

0

!!4ø,30 ' ! 14ø20 '

20 ø 20'

20 ø ,30'

20 ø 40'

20* 50'

21000 ß !1,3o40 ß

Fig. 9b. Inversion solution with rotations of the magnetization unit vector in the central area outlined in Figure 5. As in Figure 9a, two annihilators were added to this solution and heavy lines indicate the outlines of rotated bodies shown in Figure 8a. The location of the strong negative magnetizations over the block of highly oblique lineations leads us to prefer the solution with rotations to that without.

discontinuity southward to the Easter microplate. Figure 13 shows side scan imagery at two places along this segment. At 21øS, the ridge axis has an overall orientation of 008 ø while individual left-stepping faults and summit. calderas are oriented 011 o. At 22ø30'S, the ridge axis and individual faults are oriented 002 ø to 004 ø . This difference in the ridge axis orientation can be explained by north-south compression in the Nazca plate that decreases with distance from the Easter microplate.

Analysis by Sinton et aL[ 1991] of glass samples collected on the axis of the EPR t¾om the Garrett Transfc)rm to the Easter

microplate indicate that the discontinuity at 20ø40'S also con'esponds to a boundary in mantle source composition. It is possible, therefore, that while the offset of the feature may be controlled by Easter microplate tectonics, its origin is deep seated. This cannot be confidently stated, however, since geochemical analyses are limited at present to very young rocks, and it is therefore impossible to determine whether or not the boundary in mantle source composition has existed for the 2 to 3 m.y. life of the offset.

In addition, Sinton et al. [1991] found that the northern segment has the fractionation signature of a propagating rift,

supporting the idea that the current propagation event is occuning on this segment.. The highly differentiated nature of samples collected on ridge tips abandoned by both the northern and southern segments is further evidence that both segments have propagated in the past 0.1 m.y (S. Smaglik, personal communication, 1991).

Local Features

The results of this study further document and clarify some of the processes occurring at migrating discontinuities. Although the pattern of magnetization intensities indicates high intensities in the discordant zone as has been previously suggested and observed for other discontinuities [Sempdrd and Macdonald,, 1986; Pertain and Macdonald, 1990], the pattern is discontinuous and does not clearly indicate the path of the discontinuity as the pattern does at 11 ø45'N [ Perram and Macdonald, 1990] and 9øN [Carbotte and Macdonald, 1992] on the EPR. This is one piece of evidence suggesting an extremely complicated fine scale history for this discontinuity in which conditions at the discontinuity have varied a great deal over short time intervals.

PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS 13,847

0.96 Ma 0.94 Ma 0.92 Ma 0.90 Ma 0.89 Ma

Fig. 10. A model showing one possible origin for the overlapping Jaramillo isochron on the west flank of the rise axis. This model and

those of Figures 11 and 12 were developed using the program of Wilson [1990]. In all three figures, bold lines indicate the locations of the ridge axes immediately prior to the time shown. Lines with arrows indicate the new path taken by the ridge tip at the time shown. The diagram for 0.96 Ma represents a time after the beginning of the Jaramillo when there was litfie overlap of the ridge segments. This is followed (0.92 Ma) by rapid propagation of both segments causing a large overlap. At the end of the Jaramillo the northern ridge cuts outside of itself and the southern ridge cuts inside of itself. At 0.89 Ma there is again little or no overlap of the ridge segments.

.:

a Ma 0.68 Ma

est flank

iilll •'/l•--'•:•'••:• •tt oblique lineations Ill[/t't'[•'•••l [• ["•t and negative till/[II•'•[;--•[ ' '•'••l / t•l'l magnetization

0.56-Ma 0.40 Ma

Oblique lineations and crustal rotation. The rotations present on the west flank and possibly in the current overlap zone are features which have been predicted for both OSCs [ Macdonald et al., 1984] and propagators [McKenzie, 1986; Wilson, 1990]. From the models of these and other workers, there are two basic

modes of deformation or rotation associated with overlapping and/or migrating discontinuities. First, crust transferred by the propagating segment from one plate to the other is sheared and/or rotated during transfer as it passes through the region of overlap. The oblique lineations of the shear zone at the Galapagos 95ø30'W propagator [Hey et al., 1986] are an example of this type of deformation. It has been suggested that this type of deformation occurs in one of two ways. In the deck-of-cards scenario proposed by McKenzie [1986] shearing occurs along multiple faults oriented in the direction of relative plate motion. In the bookshelf faulting mechanism shearing occurs between blocks as they are rotated through the shear zone [Kleinrock and Hey, 1989]. Second, crust which remains undeformed within an overlap region (whether it was formed inside or outside that region) for a period of time will be rotated. This mode is similar to the rotation of microplates in the models of Schouten et al. [1988] and Engeln et al. [1988].

Because of the magnetic evidence for rotation, deck-of-cards shearing is not a viable mechanism for the development of the oblique lineations we observe in the present study area. We believe that the west flank zone of oblique lineations represents bookshelf faulting in a shear zone and/or block rotation in an overlap region. Distinguishing between the two mechanisms is a matter of determining the location of shear zones. With bookshelf faulting, shear zones should occur within the area of rotation, parallel or subparallel to rotated abyssal hills. With block rotation, shear zones should occur outside of the rotated

area. In the present study area, there are no high-resolution data for the west flank area which would enable identification of shear

zones. In the area of the most highly rotated fabric at the present offset, Deep Tow data show little talus [Macdonald et al., 1988] suggesting that shearing has occurred outside of this near-axis area causing block rotation.

Kinematic models of propagating spreading centers [McKenzie, 1986; Wilson, 1990] include a relatively continuous zone of oblique lineations reflecting transfer of crust from one plate to the other. These lineations are observed at both the Galapagos 95ø30'W [Hey et al., 1986] and 87ø30'W [Perram and Macdonald, 1993] propagators. If the oblique lineations indicate bookshelf faulting as suggested by Kleinrock and Hey [1989], then it is reasonable to predict that rotations will be associated with a steadily migrating discontinuity. Although the 20ø40'S dueling propagator has not migrated steadily over the past 1-2 m.y., it is possible that it has experienced periods of steady migration during that time span. Examples are the 0.08-m.y. periods of steady propagation that we interpret to have occurred during the transfer of the west flank rotated area from the Nazca to the Pacific plate (Figure 12). The OSC at 11ø45'N on the EPR

Fig. 11. A model showing one possible origin for the rotated blocks on the west flank of the rise axis. Shaded areas are positively magnetized crust. Initially (0.84-0.76 Ma), negatively magnetized pre-Brunhes crust is transferred into the zone of overlap. The ridges then connect through (0.68 Ma), and this is followed by resumed southward propagation of the northern segment (0.56 Ma), again trapping and rotating the negatively magnetized block. The southern ridge segment then retreats by cutting inside of itself, rafting the rotated block to the west flanlc This is followed by northward propagation of the southern segment (0.40 Ma).

13,848 PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS

a

rotation of several abandoned overlap basins. In this case, whole overlap basins have been rafted onto the west flank of the ridge axis. These studies indicate that rotations detectable with surface

magnetics and SeaMARC II side scan sonar can generally be expected to occur where steady migration has occurred and the ridge tip abandonment interval is large or where significant migration has not occurred but entire overlap basins have been rafted onto the flank of the rise axis.

Fine scale processes. Although the models we present for the origin of magnetic discordant zone features are only examples of many possible origins, there are aspects of the models which we believe are necessary in any model for the origin of these features. First, there is considerable evidence for great variation in the amount of overlap through time. Macdonald et al..[ 1988] recognized this in the near-axis region, and this study shows that this variation has been a pattern for the past ! m.y. In addition, as was shown by Macdonald et al. [ 1988], it is clear that both the southern and northern ridge segments have propagated and both have cut inside and outside of themselves at times in their

a 0.02 Ma

Fig. 12. A model showing the history of the discontinuity for the past 0.09 m.y. Ridges identification is from Macdonald et al. [1988]. Prior to 0.10 Ma there is a large amount of overlap and extension near the tip of the southern ridge segment forms a hole. This is followed by the outward cutting of the northern ridge segment (ridge 4) and the inward cutting of the southern ridge segment to form ridge 2 (0.06 Ma).

[Perram and Macdonald, 1990] appears to have migrated steadily southward over the past 0.2 m.y. with no evidence of rotations. There are two logical reasons for this. First, because the overlap at this OSC is large when compared to the offset and the propagation rate has been fast, the rotation should not be large. Second, there has been at least three ridge tip abandonment events during the past 0.2 m.y. so that any crustal blocks or slivers that may have been rotated have been dissected and overprinted by more recent flows.

Block rotations of crust trapped in an overlap zone may occur at discontinuities where there is not significant or steady migration. As spreading occurs on two overlapping ridge tips, the block of crust within the overlap will rotate. If the two ridges do not connect with one another, rafting the entire block onto one flank, shear zones should eventually develop at the ends of the overlap. In a recent study of the OSC at 9øN on the EPR, Carbotte and Macdonald [1992] find magnetic evidence for the

overlapping daramillo

west flank

oblique lineations and negative

magnetization

Present

Fig. 12. (continued) The southern ridge then cuts to the outside of the hole to form ridge 1' and the northern ridge cuts inside of itself to form ridge 3 (0.02 Ma). Finally, the southern ridge cuts slightly to the inside of itseft to its present location at ridge 1.

PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS 13,849

Fig. 13. Side scan sonar imagery of the axis of the EPR with interpretations of the imagery at (a) 21øS and (b) 22ø30'S. Faults and summit calderas are oriented 011 ø at 21øS and 002 ø to 004 ø at 22ø30'S. This difference reflects an increase in

north-south Nazca plate compression toward the the Easter microplate.

history. Because we have little constraint on the relative ages of ridges older than 0.1 m.y., we cannot determine any patterns of preferential inward or outward cutting of the ridge tips.

Our magnetization distribution solutions indicate a transfer of crust from the Nazca plate to the Pacific plate in at least two distinct cases. Transferred Jaramillb age crust causes a gap in the Jaramillo isochron on the east flank and a double Jaramillo

isochron on the west flank. Transferred Matuyama a•e crust causes a narrow east flank Brunhes-Jaramillo isochron and

negative anomalies and magnetizations embedded within the Brunhes. It is possible that crust has been transferred irl the most recent cycle of dueling, but a lack of reversals makes it difficult to identify a transfer if it has occurred. A continuous transfer of crust .from Nazca to Pacific is what would be expected with steady southward propagation. However, the seemingly episodic transfer of crust observed in this area is more consistent with the

dueling propagation and net southward migration which has occurked at this discontinuity over the past 1-2 m.y.

By examining magnetic and tectonic features in this area it has been possible to identify a cycle of overlap variation occurring at different times in the history of the discontinuity. Our interpretation of the overlapping Jaramillo isochron on the west flank of the ridge axis (Figuke 10) indicates a cycle of little or no overlap followed by rapid •ropagation of the ridge segments causing greater than 50 km of overlap and a return to a small amount of overlap. This cycle occurs almost entirely within the Jaramillo normal epoch and is estimated to be approximately 0.06 m.y. in length. This cycle is similar in pattern and time scale to the cycle interpreted by Macdonald et al. [ 1988] to have occurred in the past 0.08 m.y. and which we model here as having occurred over the past 0.1 m.y. The variation in overlap interpreted to have occurred in association with the formation of the west flank area of rotation also shows a cyclic pattern with

one cycle occurring over a period of-0.16 m.y. While the pattern of overlap variation may be consistent through the history of the discontinuity, the time scale of this variatiori is not consistent.

In the dueling propagation cycles discussed above, changes in the amount of overlap and changes in the position of the offset are accomplished through abandonment of ridge tips by inward or outward cutting of new ridges. In this manner it shoold be possible at any given time to find abandoned ridge tips inside and/or outside the active region of overlap as well as to the north and south of the location of the active offset. Although our model for the history of this discontinuity is nonunlque, the features in the area were most easily modeled with time intervals for ridge tip abandonment ranging between 0.01 and 0.08 m.y.

CONCLUSIONS

The major conclusions of this study are as follows: 1. We confirm that there has been consistent asymmetric

spreading in the study area for the past 2 m.y. with faster spreading to the east. This asymmetry may be related to the shallowness of west flank crust relative to east flank crust of the

same age.

2. The discontinuity of the EPR at 20 ø40'S has behaved as a dueling propagator for at least the last 1 m.y. of its history. It has experienced southward migration during that time period. This migration has slowed and possibly stopped during the Brunhes while dueling has continued.

3. The offset of the discontinuity has remained constant while the southern ridge has become rotated in its orientation. This rotation increases to the south and is likely the result of Nazca plate compression north of the Easter microplate.

4. Important fine scale processes occurring at this

13,850 PERRAM ET AL.: EAST PACIFIC RISE 20ø40'S, EVIDENCE FOR CRUSTAL ROTATIONS

discontinuity over the past 1-2 m.y. are large variations in overlap, rapid propagation of both ridges, inward and outward cutting and abandonment of ridge tips, and episodic transfer of crust from the Nazca plate to the Pacific plate.

5. The broad cross-strike oriented negative anorr;aly within the Brunhes anomaly is caused by high-intensity magnetizations associated with the tectonically disturbed areas in the path of the dueling propagators. However, these high magnetization intensities do not clearly indicate the path of the discontinuity as they do at other discontinuities along the EPR.

6. The bull's-eye-shaped negative anomaly within the central anomaly is due to the presence of highly rotated negatively magnetized pre-Brunhes Nazca plate crust within Brunhes age Pacific plate crust. The rotation of this crust appears to have been accomplished by bookshelf faulting or block rotation within the overlap zone.

Acknowledgments. We gratefully acknowledge the assistance of Captain Hayes and the crew of the R/V Moana Wave, the SeaMARC II group led by Sandy Shor, and cruise participants Rachel Haymon, Steve Miller, Patricio Goyez, Norm Brown, Suzanne Carbotte, Brian Cousens, and Jeff Severinghaus. Steve Miller provided assistance with the magnetic modeling, and Antoinette Padgett drafted many of the figures. Discussions with Suzanne Carbotte, Suzanne Smaglik, and Doug Wilson and the thoughtful comments of an anonymous reviewer improved this work. We thank the National Science Foundation for its support under grant OCE86-09706.

REFERENCES

Carbotte, S. M., and K. C. Macdonald, East Pacific Rise 8ø-10ø30'N: Evolution of ridge segments and discontinuities from SeaMARC II and three-dimensional magnetic studies, J. Geophys. Res., 97, 6959- 6982, 1992.

Cochran, J. R., Variations in subsidence rates along intermediate and fast spreading mid-ocean ridges, Geophys. J. R. Astron. Soc., 87, 421-454, 1986.

Cormier, M.-H., K. C. Macdonald, and S. P. Miller, Dueling propagating ridges on the East Pacific Rise near 20ø40'S: A gravity study (abstract), Eos Trans AGU, 69, 1474, 1988.

Davis, E. E., and J. L. Karsten, On the cause of the asymmetric distribution of seamounts about the Juan de Fuca ridge: Ridge-crest migration over a heterogeneous asthenosphere, Earth Planet. Sci. Lett., 79, 385-396, 1986.

DeMets, C., R. G. Gordon, D. F. Argus, and S. Stein, Current plate motions, Geophys. J. Int., I01, 425-478, 1990.

Engeln, J. F., S. Stein, J. Werner, and R. G. Gordon, Microplate and shear zone models for oceanic spreading center reorganizations, J Geophys. Res., 93, 2839-2856, 1988.

Haftand, W.B., A.V. Cox, P.G. Llwellyn, C.A.G. Pickton, A.G. Smith, and R. Walters, A Geological Time Scale, Cambridge University Press, New York, 1982.

Hey, R.N., M. C. Kleinrock, S. P. Miller, T. M. Atwater, and R. C. Searle, Sea Beam/Deep-Tow investigation of an active oceanic propagating rift system, Galapagos 95.5øW, J. Geophys. Res., 91, 3369-3393, 1986.

IAGA Division I Working Group 1, International geomagnetic reference field revision 1987, IAGA News, 26, 87-92, 1987.

Kleinrock, M. C., and R. N. Hey, Migrating transform zone and lithospheric transfer at the Galapagos 95.5øW propagator, J. Geophys. Res., 94, 13,859-13,878, 1989.

Klitgord, K.D., Sea-floor spreading: The central anomaly magnetization high, Earth Planet. Sci. Lett., 29,201-209, 1976.

Lonsdale, P., Segmentation of the Pacific-Nazca spreading center, 1ø- 20øS, J. Geophys. Res., 94, 12,197-12,226, 1989.

Macdonald, K. C., S. P. Miller, S. P. Huestis, and F. N. Spiess Three- dimensional modeling of a magnetic reversal boundary from inversion of deep-tow measurements, J. Geophys. Res., 85, 3670- 3680, 1980.

Macdonald, K. C., J.-C. Semp6r6, and P. J. Fox, East Pacific Rise from Siqueiros to Orozco fracture zones: Along-strike continuity of axial

neovolcanic zone and structure and evolution of overlapping spreading centers, J. Geophys. Res., 89, 6049-6069, 1984.

Macdonald, K. C., R. M. Haymon, S. P. Miller, J.-C. Semp6r•, and P. J. Fox, Deep-Tow and Sea Beam studies of dueling propagating ridges on the East Pacific Rise near 20ø40'S, J. Geophys Res., 93, 2875- 2898, 1988.

McKenzie, D. P., The geometry of propagating rifts, Earth Planet. Sci. Lett., 77, 176-186, 1986.

Miller, S. P., and R. N. Hey, Three-dimensional magnetic modeling of a propagating rift, Galapagos, 95ø30'W, J. Geophys. Res., 91, 3395- 3406, 1986.

Naar, D. F., and R. N. Hey, Recent Pacific-Easter-Nazca plate motions, in Evolution of Mid Ocean Ridges, Geophys. Monogr. Ser., vol. 57, edited by J.M. Sinton, pp. 9-30, AGU Washington, D.C., 1989.

Naar, D.F., and R.N. Hey, Tectonic evolution of the Easter microplate, J. Geophys. Res., 96, 7961-7993, 1991.

Parker, R. L., and S. P. Huestis, The inversion of magnetic anomalies in the presence of topography, J. Geophys. Res., 79, 1587-1593, 1974.

Perram, L. J., and K. C. Macdonald, A one-million-year history of the 11ø45'N East Pacific Rise discontinuity, J. Geophys. Res., 95, 21363-21381, 1990.

Perram, L. J., and K. C. Macdonald, An overlapping propagating spreading center at 87ø30'W on the Galapagos spreading center, Earth Planet. Sci. Lett., in press, 1993.

Prince, R. A. and D. W. Forsyth, A simple objective method for minimizing crossover errors in marine gravity data, Geophysics, 49, 1070-1083, 1988.

Rea, D. K., Asymmetric sea-floor spreading and a nontransform axis offset: The East Pacific Rise 20øS survey area, Geol. Soc. Am, Bull., 89, 836-844, 1978.

Rusby, R.I., and R.C. Searle, Intraplate thrusting near the Easter microplate (abstract), Eos Trans. AGU, 72(44), Fall Meeting suppl., 510, 1991.

Schouten, H., and K. McCamy, Filtering marine magnetic anomalies, J. Geophys. Res., 77, 7089-7099, 1972.

Schouten, H., D. G. Gallo, and K. D. Klitgord, Microplate kinematics of the second order (abstract), Eos Trans AGU, 69, 488, 1988.

Searle, R. C., R. I. Rusby, J. Engeln, R. N. Hey, J. Zukin, P.M. Hunter, T. P. LeBas, H.-J. Hoffman, and R. Livermore, Comprehensive sonar imaging of the Easter microplate, Nature, 341, 701-705, 1989.

Semp6r6, J.-C., High-magnetization zones near spreading center discontinuities, Earth Planet. Sci. Lett., 107,389-405, 1991.

Semp6r•, J.-C., and K. C. Macdonald, Overlapping spreading centers: implications from crack growth simulation by the displacement discontinuity methods, Tectonics, 5, 151-163, 1986.

Semp•r•, J.-C., K. C. Macdonald, and S. P. Miller, Overlapping spreading centres: 3-D inversion of the magnetic field at 9ø03'N on the East Pacific Rise, Geophys. J. R. Astron. Soc., 79, 799-811, 1984.

Sinton, J. M., S. M. Smaglik, J. J. Mahoney, and K. C. Macdonald, Magmatic processes at superfast spreading ridges: Glass compositional variations along the East Pacific Rise, 13-23øS, J. Geophys. Res., 96, 6133-6156, 1991.

Stein, S., H. J. Melosh, and J. B. Minster, Ridge migration and asymmetric seafloor spreading, Earth Planet. Sci. Lett., 36, 51-62, 1977.

Tighe, S. A., P. J. Fox, R. S. Detrick, and R. C. Tyce, A synthesis of Sea Beam data from the East Pacific Rise between the Orozco and

Siqueiros transforms (abstract) Eos Trans. AGU, 68, 1505, 1987. Wilson, D. S., Kinematics of overlapping rift propagation with cyclic

rift failure, Earth Planet. $ci. Lett., 96, 384-392, 1990.

M.-H. Cormier and K.C. Macdonald, Department of Geological Sciences, University of California, Santa Barbara, CA 93106.

L.J. Penvenne, Triton Technology, Inc., 125 Westridge Drive, Watsonville, CA 95076.

(Received July 15, 1991; revised November 16, 1992;

accepted November 29, 1992.)