A variety of seismic experiments indicate that seismic velocities in the uppermost oceanic crust (layer 2A) increase very rapidly with age, nearly doubling in 10 my or less, but only one measurement was made in situ, by an oblique seismic experiment in DSDP/ODP Hole 504B. A VSP acquired in 11 Ma crust at ODP Site 1243 yields a velocity of 4.21 ± 0.08 km s−1, which is consistent with previous observations. The absence of a vertical velocity gradient at Site 1243 is consistent with the hypothesis that layer 2A is modified by a process of progressive sealing of cracks from the bottom up. Application of an asperity deformation model suggests that the area of contact across cracks in layer 2A increases by a factor of ∼100 between 1 and 10 my.
 Site 1243 is located in the vicinity of DSDP Site 852, near 5°18′N, 110°04.5′W in the eastern equatorial Pacific. Because this is an equatorial site, the age of the seafloor cannot be estimated from magnetic lineations. Based on subsidence and a full spreading rate of 141 mm/a, the age of the crust at the site is 10–12 Ma [Orcutt et al., 2003]. Two holes were drilled at Site 1243. Hole 1243A is a legacy hole, intended for later installation of a digital broadband seismic observatory. Hole 1243B was drilled 600 m to the east of Hole 1243A to characterize basement at the site and to establish a reference section in 10–12 Ma Pacific crust.
 Hole 1243B was drilled to a depth of 195.3 mbsf (meters below seafloor), and penetrated basement to a depth of 87.1 meters. Of eight lithologic units, five are comprised of tholeiitic pillow basalt; Unit 2 is a thin limestone (based on a single piece of limestone in the core) and Unit 4 is alkalic basalt. Unit 8 is a glass-rich drilling breccia. There is no evidence of massive flow units or intrusive rocks in the recovered core.
 Physical properties measurements were made on 20 basalt samples: bulk densities range from 2.52 to 2.82 kg m−3, with a mean of 2.69 ± 0.02 kg m−3' porosities range from 4% to 17%, with a mean of 7.7 ± 0.7%' and P-wave velocities, measured at bench pressure, range from 4.3 to 5.7 km s−1, with an average of 5.26 ± 0.08 km s−1. In addition, a sonic log was obtained from the upper part of the basement section (core depth 110 to 149 mbsf) [Orcutt et al., 2003]. The mean sonic velocity over this interval is 4.73 km s−1.
3. Vertical Seismic Profile
 The Schlumberger Well Seismic Tool (WST) was used to obtain a vertical seismic profile (VSP) through the basement section of Hole 1243B from 110.7 to 180.0 mbsf [Orcutt et al., 2003]. The source was a Sodera 200 cubic inch airgun at a depth of 2 m below the sea surface, and offset 55.8 m from the location of the borehole. Shot break records were recorded by a hydrophone at the source. To obtain the VSP, the WST was lowered to the bottom of the borehole, and the caliper arms were extended to clamp the instrument to the borehole wall. The cable was then slacked off 5 meters, to decouple the logging cable from the instrument. Several shots were recorded with a sampling interval of 1 ms, and the process was repeated at several stations as the instrument was drawn up the hole. Except for the top and bottom stations, records were taken at ten-meter intervals from 180 to 110.7 mbsf. The depth and the number of shots recorded at each station are listed in Table 1. The stacked traces are shown in Figure 2. The dominant frequency of these records is near 100 Hz. It should be noted that this frequency is higher than the frequencies recorded in seismic refraction data.
Table 1. Summary of WST Data
2646.02 ± 0.09
2649.02 ± 0.13
2651.28 ± 0.07
2653.57 ± 0.16
2655.92 ± 0.08
2658.57 ± 0.16
2661.28 ± 0.07
2664.85 ± 0.07
4. Data Analysis
 Traveltimes were estimated from the lags between the stacked shot break records recorded at the source and the stacked signals recorded at the WST, using a modified version of a cross-correlation method developed by Gangi and Fairborn , who showed that the cross correlation Φn(τ) in the neighborhood of the correct lag τn is a quadratic function
In this case, an initial estimate of the lag τ was made by inspection, and a refined estimate was obtained by fitting a quadratic function to the computed cross correlation versus lag function over a five-lag interval, usually from (τ − 2) to (τ + 2). The correct lag was then determined by setting the first derivative of the lag function Φn(τ − τn) ≡ 0, and solving for τn. An example of the matched source and WST signals is shown in Figure 3. Traveltimes and associated standard errors ranging from 0.07 to 0.16 ms are listed in Table 1. The traveltimes and errors are comparable to those determined by a threshold method [Orcutt et al., 2003].
 Traveltimes are shown as a function of depth in Figure 4. Orcutt et al.  noted that the arrival at station 110.7 is apparently early, and the interval velocity between 110.7 and 115.7 mbsf is 1.67 km/s – more characteristic of sediments than of even the lowest velocity basement rocks. For that reason, data from the first station were not used in this analysis. The P-wave velocity over the interval from 115.7 to 180 mbsf was estimated by linear regression, and found to be 4.03 ± 0.11 km s−1, while the velocity over the sonic log interval (approximately 115.7 to 155.7 mbsf) is 4.21 ± 0.08 km s−1.
5. Discussion and Conclusions
 The data from Hole 1243B are included with the data compiled by Grevemeyer et al.  in Figure 1. These data extend the maximum age of sites where layer 2A velocities have been measured from 8 to 11 Ma (the data point at 10 Ma represents the median for ages ranging from 5 to 20 Ma from Carlson ). The velocities measured at Site 1243 are consistent with previous observations and inferences regarding the evolution of layer 2A in several respects:
 1) The Hole 1243B sonic log and WST velocities are, as expected, appreciably lower than the velocities measured in the laboratory [Orcutt et al., 2003]. This observation is consistent with early evidence that layer 2A is populated by cracks that cause seismic velocities in the formation to be dramatically lower than the measured velocities in the rocks that comprise the uppermost crust [e.g., Hyndman and Drury, 1976]. Also worth noting is the fact that the sonic velocity is higher than the WST velocity. This difference suggests that the sonic velocity represents a time average while the WST velocity represents a long wave approximation [e.g., Backus, 1962]. Given that the frequency of the sonic signal is of the order of 104 Hz, the sonic wavelength is ∼40 cm – near the characteristic dimension of a basalt pillow - while the wavelength of the seismic signal is near 400 m.
 2) The WST velocities from Hole 1243B are entirely consistent with the velocities in layer 2A obtained by other methods, and with the inference that layer 2A velocities rise very rapidly during the first 6 to 10 my after formation of the crust, and change little, if at all, thereafter [Grevemeyer and Weigel, 1996, 1997; Carlson, 1998; Grevemeyer et al., 1999]. The velocities shown in Figure 1 are consistent with a continuing, gradual increase of upper crustal velocities from 6 and 11 Ma, but in view of the fact that the uncertainties in the measured velocities are typically ∼0.1 km s−1, there is no compelling evidence for a systematic increase of seismic velocities in layer 2A at ages greater than 6 Ma.
 3) On the other hand, both the average sonic log and WST velocities measured in Hole 1243B are significantly lower than the velocity that is thought to be typical of layer 2B, about 5.2 km s−1. This observation is consistent with the inference that layer 2A persists as a thin, low-velocity cap on the extrusive oceanic crust [Carlson, 1998].
 On the whole, the results of the WST survey of the uppermost basement section in Hole 1243B are consistent with recent inferences regarding the evolution of layer 2A. It is perhaps worth exploring as well what the pattern of increasing velocity with age shown in Figure 1 has to tell us about the alteration process. It is widely agreed that the increase of velocity with age in layer 2A is caused by mechanical collapse and progressive infilling of cracks in the formation by the mineral products of hydrothermal alteration, as originally proposed by Houtz and Ewing . Cracks are commonly modeled as ellipsoidal inclusions [e.g., Kuster and Toksöz, 1974; O'Connell and Budiansky, 1974]. Wilkens et al.  used such a model to demonstrate that the range of velocities observed in layer 2A can be accounted for by filling only the very thinnest cracks, with little or no reduction of bulk porosity. Alternatively, asperity deformation models [e.g., Gangi, 1978] also account for the influence of cracks on the seismic properties of rocks. Vera and Mutter  suggested that progressive stiffening of asperities under increasing overburden pressures can account for velocity gradients in the oceanic crust. Because alteration products modify the distribution of asperities that populate crack walls, an asperity model can also account for the observed variation of velocity with age in layer 2A.
 In the “bed of nails” asperity deformation model [Gangi, 1978], the seismic velocity is given by
where Pe is the effective pressure, Pi is the “initial” pressure, which accounts for the fact that some asperities are in contact when Pe = 0, and the exponent m is related to the distribution of asperity heights. Overburden pressures are supported in part by the asperities and partly by pore pressure acting on the walls of the cracks where the asperities are not in contact. Gangi and Carlson  have shown that Pe = Pc + (1 − Af)Pp = Pc − Pp + AfPp. Here Pc is the confining pressure, Pp is the pore pressure, and Af is the fractional area of contact of the asperities across the cracks. In the uppermost crust Pc ∼ Pp, which is constant. The velocity then varies with Af, which increases as the cracks are filled by alteration products. If we assume the Af is simply proportional to age (the simplest possible model), the variation of velocity with age can be written as
A fit of this simple model to the layer 2A velocity versus age data is included in Figure 1. The model accounts for the data quite well, with Vo ∼ 2.4 km s−1 and b ∼ 10, with an RMS error of 0.28 km/s and R2 = 0.93. This result suggests that Af increases by a factor of about 100 as the crust ages from 1 to 10 Ma.
 On the whole, the results of this study indicate that the seismic velocity in the uppermost crust at ODP Site 1243 are consistent with previous results that indicate that velocities change rapidly as a result of closure or infilling of thin cracks by alteration in the first few million years after formation of the crust and remain more or less constant after about 10 million years.
 This research used samples and/or data provided by the Ocean Drilling Program (ODP). ODP is sponsored by the U.S. National Science Foundation (NSF) and participating countries under management of Joint Oceanographic Institutions (JOI), Inc. Funding for this research was provided by a grant from the US Science Support Program. I wish to thank the Captain and crew of Joides Resolution. I am particularly indebted to the Ocean Drilling Program and to Kerry Swain and Arno Buysch, who acquired the data. I also wish to thank Paul Johnson and an anonymous reviewer for their helpful comments.