JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 87, NO. B8, PAGES 6650-6656, AUGUST 10, 1982 Microcrack Studies of Basalts From the Iceland ResearchDrilling Project BART J. KOWALLIS,EVELYN A. ROELOFFS, AND HERBERTF. WANG Department of Geology and Geophysics,Universityof Wisconsin-Madison Madison, Wisconsin 53706 Microcavities in basalticcore from the IRDP borehole in Reydarfj6rdur, Iceland, can be characterized by using direct observationswith the scanningelectron microscope(SEM) and inversion of velocity data. In the Iceland basalts, neither method provides an accurate, complete picture of the crack aspect ratio spectrum, but by using both SEM and inversion data a complete spectrum may be obtained. SEM observationsprovideinformation on high aspectratio poresand cracks(>0.005) that cannot be separatedby velocity inversiontheory, while velocity inversiontheory providesinformation on low aspect ratio cracks(< 0.005) that are too narrow to be seenwith the SEM. Alteration rims along cracks,blunt ends, and aspectratios which imply closurepressuresgreaterthan the vertical stressall provide evidence that cracks are open in situ. Cracks coincident with grain boundariesare abundant in the basalt flows and suggestan origin early in the rock's history, probably during cooling. Crack healing and alteration are observed to increase with depth, causing a shift in crack shape distribution toward higher aspect ratios. INTRODUCTION oceanic crustal basalts. Although the basalts in Iceland are not really typicalof mid-oceanridgebasaltssincethey formed Microcavities are microscopicopeningsfound in nearly all subaerially,they do contain a record of eruptiveprocessesat rocks. Those of equal or nearly equal dimensionsare called a spreadingaxis and are well exposedand easilyaccessiblefor micropores,while microcavitieswith one dimensionsmaller or study [Gibson, 1979]. larger than the other two are called microcracks. Microcavities Several theoretical models of the effect of cracks on significantlyinfluencethe elasticpropertiesof rocks as well as elastic propertiespresently exist [Brace, 1965; Walsh, 1965; their electrical,transport, thermal, and strengthcharacteristics O'Connell and Budiansky, 1974; Toks6z et al., 1976]. Hadley [Sprunt and Brace, 1974]. Microcracks have also been used [1976] showedthat although agreementbetweenlaboratory as a means of interpreting the geologic and tectonic history data and these theories is generally good, the problem is of lunar [Simmons et al., 1975; Richter et al., 1976] and underconstrained because of a lack of data on actual observed terrestrial rocks [Simmons and Richter, 1976; Richter and pore and crack geometries. Hadley used SEM observations Simmons, 1977]. on the Westerly granite to provide the needed additional Brace et al. [1972] developed a technique for observing geometricalconstraints. We use crack aspectratio data taken microcavities directly by first preparing a finely polished rock from SEM photos to produce a crack aspectratio spectrum surface and then ion thinning the sampleto produce a surface for each sampleand comparethesespectrawith crack aspect free of damage causedby grinding. This surface then can be ratio spectra generatedby inverting ultrasonic velocity data coated and observed using the scanningelectron microscope usinga techniquedescribedby Chengand ToksOz[1979]. Our (SEM). The preparationof specialrock thin sections,called velocity data compare favorably to velocity data from similar crack sections, allows observation with both the SEM and the sectionsof the coremeasuredby Christensenand Wilkens[this petrographic microscope[Simmonset al., 1975; Simmons and issue]at the Universityof Washington. Richter, 1976]. Crack sectionsare thicker (--100 microns) than normal thin sectionsand are cut using a slow speedsaw CORE DESCRIPTION after which they are polished by hand. This processprevents damage to the section during preparation and thus eliminates The IRDP core consists of both basalt flows and dikes. most cracksthat are createdduring thin sectionpreparation. Complete flows can be identified by their red, brecciated, Sprunt and Brace [1974] observed microcavities in a wide scoriaceous tops and bottoms, whereas other flows are variety of crystalline rocks and gave some convincing criteria identified by their finer-grained vesicular and weathered to support the idea that microcavitiesare presentin the crust. texturesin comparisonwith the dike units. Four sampleswere Wang and Simmons [1978] showed, however, using gabbro chosenfor study from differentdepthsand units in the cored from 5 km deep in the Michigan Basin, that not all crustal sequence.The sampleswill be referred to in the text by their rocks contain open microcracks at depth. It seemsthat in depth in metersfrom the top of the hole. predictionof crustalproperties(e.g., seismicvelocity,thermal We studied two samplesfrom the central part of basalt conductivity, or electricalconductivity) for any given area, flows (233.2 and 1494.1), one samplefrom a flow top we need to know something about not only rock type but (1487.9),and one from a dike (1813.9). Samples233.2 and also the presenceor absenceof microcavities. The core from 1494.1 are light gray to greenish-gray,fine-grainedbasalts the Iceland ResearchDrilling Project (IRDP) provided an with occasionalvesiclesthat are usually filled with chlorite, opportunity to examine the distribution of microcavities in smectite,and zeolites. The flow top sample (1487.9) is brick red and brecciatedwith considerableepidotefilling or Copyright ¸ 1982 by the American GeophysicalUnion. partially filling vugs and cracks. Sample 1813.9 from the Paper number 1B0832. 6650 0148-0227/82/001 B-0832505.00  KOWALLIS ET AL.: MICROCRACKSIN IRDP BASALTS 6651 TABLE 1. Ultrasonic Velocities as a Function of Hydrostatic Pressure Bars Confining Pressure Sample Density, Depth, m g/cm 3 Mode 50 100 250 500 750 1000 1250 1500 2000 2500 3000 4000 233.2 2.88 P 5.81 5.85 5.91 5.96 6.01 6.03 6.04 6.05 6.10 6.14 6.17 6.18 S 3.38 3.39 3.39 3.40 3.40 3.42 3.43 3.43 3.44 3.44 3.45 3.46 1487.9 2.43 P 3.28 3.50 3.83 4.24 4.37 4.58 4.67 4.76 4.81 S 2.02 2.20 2.43 2.59 2.65 2.72 2.77 2.81 2.87 1494.1 2.97 P 6.00 6.05 6.14 6.18 6.27 6.30 6.34 6.38 6.44 6.45 6.46 6.57 S 3.55 3.60 3.70 3.75 3.77 3.79 3.80 3.81 3.83 3.85 3.87 3.91 1813.9 3.03 P 5.46 5.55 5.63 5.77 6.02 6.21 6.33 6.44 6.53 6.56 6.59 6.64 S 3.42 3.46 3.53 3.64 3.79 3.89 3.95 4.01 4.06 4.13 4.19 4.21 Velocities are in km s . dike is a gray, medium-grainedbasaltwith small (1-2 mm) secondarymineral growth that has healedand sealedmost of plagioclasephenocrystsand is much fresher than the other them. With the petrographicmicroscopethese sealedcracks samples. The entire core was describedby P. T. Robinson can be seen(Figure lb). They are usuallyintergranularand and H.-U. Schmincke(unpublished information, 1979) soon may extendacrossthe entire thin section. Sincethis sample after drilling. Our descriptionsfit well with theirs. is a flow top breccia, such cracks are not unexpected. Samples233.2 and 1494.1 are both from the central part DATA of basalt flows but are quite different in the distribution of microcavities. The shallower sample (233.2) contains Ultrasonic Velocities almost no micropores but has many open microcracks. The microcracks occur along grain boundaries and within Compressional and shear wave velocities were measured individualgrains,terminatingat the grain boundaries(Figure on the basaltsas a function of hydrostatic confining pressure 2a). Intergranularcracks,thosethat traversemore than one up to 4 kbar (400 MPa) exceptfor the very porouscore from grain, are not common. Alteration can be seen along open 1487.9 m that crushedat lower pressures.One-inch-diameter microcrackswith some secondarymineral growths (mostly cores were dried at room temperature in a dessicatorfor at iron oxides)filling occasionalcrackscompletely(Figure 2b). least 1 week and then jacketed in Tygon tubing. The samples In general, however, the cracks in 233.2 are more open were not heated during drying to prevent the production and continuousthan those found in 1494.1 (Figures2c, 2d, of new cracks. The 1-MHz resonant frequency ceramic 3a, and 3b). The microcracksin the deeper core are more compressional or shear wave transducers cemented between commonly bridged and are often blunt rather than tapered. two aluminum discs were used in determining velocities. They also do not always show the extensive alteration The wave arrivals were timed with a variable delay line (oxidation) along cracksthat is found in the shallowercore containing siliconeoil. Overall accuracyof the measurements (Figure 3a). is estimated at about 2% with error mainly due to picking The core from 1813.9 m is part of a large basaltic dike preciselythe first arrival of a wave. The velocity data are and containsthe fewest microcavitiesof any of the samples. reproducedin Table 1. Those that are present are mostly isolated micropores and grain boundary cracks around plagioclaselaths (Figure 4). Petrography and Crack Geometry Comparatively little alteration or secondarymineralization is present in the sample. Microcavities were observedin all the core samplesusing Two-dimensional crack aspectratios were calculated from the petrographic microscopeand the SEM. Crack sections measured crack widths and lengths from each sample, were prepared using the technique describedby Simmons et where aspect ratio equals width/length. Crack dimensions al. [1975] and Simmons and Richter [1976]. These crack were taken directly from SEM photographs, ranging in sections are about 100 microns thick and are cut on a slow magnification from 100 to 10,000 X. In some cases the saw, polished by hand, and then ion thinned for 5-6 hours widths had to be estimated, and those narrower than about using an ionized argon beam to remove any damaged surface 0.03 microns were not observable due to the gold coating material. The samples thus prepared are free of cracks on the sample [Sprunt and Brace, 1974; Hadley, 1976]. that can be produced by regular sectioning techniques and Since most of the open cracks observed occur as grain they can be used on both the SEM and the petrographic boundary cracks or are contained within a single grain, microscope. Actual observation using the crack sections their length is limited by the size of grains. In the basalts on the petrographic microscopeis difficult with the basalts the average grain is about 0.1 mm long, while the average because of the high content of opaque minerals and the crack length observed was approximately 0.01 mm. Since small crystal size. the SEM cannot resolve cracks narrower than about 0.03 Sample 1487.9, from a flow top, contains mostly angular microns, most cracks of aspect ratio less than 0.003 were or rectangular microporesthat are often filled with epidote probably not seen and many with aspect ratios less than or zeolite (Figuresla-lc). The diameterof theseopen pores 0.005 were probably missed. We feel that most cracks with ranges from less than a micron to more than 2 mm. Open higher aspectratios (>0.005) are observableusing the SEM microcracks are uncommon in 1487.9 because of extensive in the Iceland basalts. Hadley [1976] noted an aspect ratio 6652 KOWALLIS ET AL..' MICROCRACKS IN IRDP BASALTS Fig. 1. Sample 1487.9, a flow top breccia. (a) Large pore partially filled with epidote crystals. (b) Healed intergranularcrack cutting acrosspartially altered plagioclaselaths and glassymatrix. Field of view is about I mm. (c) Typical pore structurein flow top sample. resolution limit of --- 0.001 in her study of Westerly granite, basalts it is not uncommon for large open pores, vugs, or in which the average grain size and crack length are greater cracks to be present. Because of the small area that was than in the Iceland basalts. examined with the SEM, we may have missed large open An averaging technique was used in compiling crack pores included in the measuredimmersion porosity of sample distribution spectra from the SEM photos. Several photos 233.2. In sample 1494.1, similarly from a flow center, we were taken at different magnificationsof the same area. Some do not see this discrepancybetween measured and observed microcavities could be seen at 100 X, while additional cracks porosities, probably because in the deeper flows most of and poresappearedat 300 X, 1000X, etc. The microcavities the large vugs, pores, and cracks are filled with secondary that were seen and measured at low magnification were not minerals. measured again, but the smaller and narrower cavities that could not be seen.at lower magnification were measured. Elastic Moduli These cracks measured at higher magnifications and thus smaller areas were assumedto be present over the entire area In predicting aspect ratio spectra using the Cheng and of the lower magnification photo. Usually several higher Toks6z [1979] velocity inversion method, it is necessary to magnification photos were taken and the crack distribution know the bulk and shear moduli of the solid portion of averaged over them before applying the assumption that the rock. First, we estimated the modal composition of the distribution was uniform over the area of the lower the samples as seen in thin section and then calculated magnificationphotos. The calculatedporositieswere derived bulk and shear moduli using the tables of aggregate elastic using an area average approximation which assumesthat properties compiled by Simmons and Wang [1971]. These the area of the void space seen in the photos is equal tables give both Voigt and Reuss averages for the elastic to the actual porosity [Hadley, 1976]. Good agreement constants which provide upper and lower bounding limits. was generally obtained between immersion porosities and We used the mean value of the two averages as a starting porositiescalculatedfrom the SEM photos (Table 2). The point for our calculations and varied the moduli within one exception was sample 233.2, where the porosity was the bounding limits in order to obtain a best fit of the underestimated by the SEM photo method. In the Iceland calculated velocities with the original laboratory velocities.  KOW•LIS ET •.: MICROC•CKS IN IRDP BAS•TS 6653 Fig. 2. Sample 233.2 from the central part of a basalt flow. (a) Cracks showingiron oxide alteration rims cutting through and around a grain. Light, diamond-shapedcrystalsare magnetite. The bar is 10 microns. (b) The long central crack of Figure 2a has been healedand is shownmagnified. (c) Alteration rims occur along nearly all open cracks, and they trace the outline of some that are healed or too narrow to be seen at this magnification. The white bar is 10 microns. (d) In thin sectionmany cracksare visible as seenin this plagioclasegrain. The field of view is about I mm. ....ß....... ":':':""7:. """ "'" "ß ':"'" ':' ":'"' ..... '""'.'":'::: ':':'"" ':':. -:'" "':' ':'" '":::'"'""'::" ':' .................................... '"':" '...................... "."" '"': .'""":':'"'-" ..... •.: ß • . '-'•,• '" •:: '?.. .. .'....:...•.:.:.-.....:.:..::.::...:.::.•..-:........:.•-..:.,.-::•,•::.•.::,.... ß...... ., .,•-::•:.•,--•.•.•.::. :•..:: ..::........ ::....... -,•. •.:::..::,........ ....:.:..:.....:::.::....;....:....: :.:..... :.:;:,.::::::::-..:•..........;::.::. ß .....::,.::,% ?,..,7:,:. ,,-?7': .... Fig. 3. Sample1494.1 from the centralpart of a basalt flow. (a) A line of blunt cracksremainsbetweenbridged portionsof a partially healedcrack. High aspectratio poresare also common. (b) Large crack with alterationrim and a few delicatebridges. 6654 KOWALLIS ET AL.' MICROCRACKSIN IRDP BASALTS open, the decrease in its volume with increasing pressure causes an increase in the effective moduli that is very small compared with the increasethat occurs when the pore closes completely. As a result, the inversion scheme is best suited for determining the total porosity that closesin each pressure increment. For the same reason, the inversion scheme has limited ability to apportion porosity between pores having aspect ratios greater than about 0.005 because these do not close within the experimental pressurerange. Also, pores so flat that they were closed before the first pressure increment cannot be distinguished. Uncertainties in the grain moduli and static bulk modulus of the dry rock also influence results of the inversion. The aspect ratios present in the computed spectrum are those for which closureoccursover the range where changein velocities has been measured. The closure pressure is approximately proportional to K*/o•, where K* is the static bulk modulus Fig. 4. Sample 1813.9 from a basaltic dike. Iron oxide grain with of the dry rock and cz is the aspect ratio. The constant healed cracks and a few open pores. Field of view is about 30 of proportionality dependssomewhaton the grain modulus, microns. but the main source of uncertainty is in K*, which was not measured but was taken to be close to the dynamic bulk This method of adjusting the mineral moduli is reasonable modulus of the rock at the highest experimental pressure. when one considers that up to 15% variability may occur This estimateis probably too high, especiallyat low pressures, between the Voigt and Reuss averages, plus the additional but still correct within a factor of about 3. Therefore, the error of up to 10ø70in the modal compositions. Moduli for aspectratios in the computedspectrumcould be shifted up by glassymatrix material found in the basalts were taken from this factor. the recent work of Meister et al. [1980]. Figure 5 showsthe resultsof the comparison of calculated and observed aspect ratio spectra for the four Iceland CRACK ASPECT RATIO SPECTRA basalt samples. Cheng and Toks6z [1979] did this type of comparison for the Westerly granite using SEM data from The distributions of microcavity shapes observed with Hadley [1976] and cameto the conclusionthat correspondence the SEM were compared with pore aspect ratio spectra was remarkable. The actual correlation shown by Cheng obtained by inverting ultrasonic velocity data. The inversion and Toks6z [1979] between their data and the data from is similar to the direct linear method presentedby Cheng Hadley [1976] is limited to a very narrow range of aspect and Toks6z [1979] except that the resulting linear system ratios. In the Iceland basalts the comparison is limited to is solved using a generalized inverse method [Aki and an even narrower range (Figure 6). The small grain size Richards, 1980], constrainedso that the fractional porosities in the basalts constrains, to a large extent, the length of at aspect ratios greater than 0.001 sum to the measured total the microcracks present. Since the average crack length is porosity. Cracks with aspectratios lessthan 0.001 contribute much smaller in the fine-grainedbasaltsthan it is in a typical porosityincrementsbeyondthe precisionof the total porosity granite, the resolution limit of the SEM, in terms of aspect measurement. Briefly, given a list of aspect ratios,-data for ratio, is not as good. Cracks with aspect ratios less than shear and compressionalvelocitiesas functionsof confining 0.001 were almost never observedin our samples, and many pressure, the static bulk modulus of the dry rock, and the with aspectratios lessthan 0.005 were probably missedwith elastic propertiesof the grains, the program computesthe the SEM. On the other hand, inversion of the velocity data distribution of porosity among the aspect ratios such that produces a crack spectrum that is accurate only for cracks theoretical effectivemoduli for the porous medium agree, as that close within the range of experimental pressuresused. functions of pressure,with the moduli computed from the As previouslystated, the inversiontheory cannot distinguish velocities. between cracks of different aspect ratio if the cracks do not When comparing the results of the aspect ratio inversion close. For our samples, the last cracks to close would have with directly observed crack shape distributions, it is aspect ratios around 0.005. Therefore, the discrepancies important to consider the reliability and resolution of the we see in Figure 5 between observedand calculated spectra computed aspect ratio spectrum. As long as a pore remains are due to a combination of the insensitivity of velocity TABLE 2. Porosity, Bulk Density, and Bulk Modulus Sample Immersion SEM Bulk K*, Bulk Depth, m Porosity, % Porosity, % Density Modulus 233.2 2.75 0.51 2.88 655 1487.9 10.93 13.08 2.43 310 1494.1 1.32 1.34 2.97 675 1813.9 0.36 0.37 3.03 650 Densities areing/cm 3 bulkmoduli inkbar.  KOWALLIS ET AL.: MICROCRACKS IN IRDP BASALTS 6655 SAMPLE 233.2 SAMPLE 1487.9 10-4 1o-a 1o-Z 1o-l 10-4 10 -z 10-1 7O I '70 60 60 60 _ _ 5O 5O $O- Z 4O 4O 4O 3O 3O 3O 2O - 20 20 ::, I I 1: o 10 10 o lO-4 10 -• 10 -z 10-1 1 00_ 10-s 10-z ASPECT RATIO ASPECT RATI{3 SAMPLE 1494.1 SAMPLE 1813.9 10-4 10-s 10-z 10-t 10-4 10-s 10-z 10-t 7O 70 L _ •o •0 _ _ •o •0 _ _ _ 4O 4O _ _ 30- 30 3O 20 20 2O 10 0 10 -4 10-s 10-Z 10 -• 1 o 10 oi 10-4 10-S 10-Z 1O- t ]12ø ASPECT RATIO d ASPECT RATIO Fig. 5. Comparisonsof aspectratio spectrafrom SEM observations(solid bars) and spectracalculatedfrom the generalizedlinear inversion(openbars). Plots are of aspectratio versuspercentof the total porosityfound at a particularaspectratio. The percentporositiesshownat particularaspectratios representthe porosity of all cracks lessthan or equal to that aspectratio above the next lowest aspectratio slot. inversion theory to pores or cracks that do not close within microcavities are open in situ. These observations include the experimental pressurerange and the fact that many low alteration rims along cracks and around pores (Figures2a, aspectratio cracks cannot be seenwith the SEM becausethey 2c, and 3b), suggestingthe presenceand transport of fluids are too narrow. One can see, however, similarity between in thesecracksand pores, partially healed or bridged cracks calculated and observed spectra in the range where the two (Figures3a and3b)andpartiallyfilledpores(Figurela), blunt methodsoverlap, particularly in samples233.2 (Figure 5a) terminationsof cracks(Figuresl c, 3a, and 4) rather than and 1494.1(Figure5c). the taperedtype of crack termination that is typical of newly Sample 233.2 is unusual in that the SEM spectrum gives formed cracks, and delicate mineral growths that would not relatively higher fractional porosities of low aspect ratio have survivedany secondarycrackingprocess.Also, the SEM cracks than does the inversion spectrum. This sample is the crack aspect ratio spectra indicate that most of the cracks same one for which the porosity was underestimatedby SEM presentin the samplesare cracksthat would closeat pressures observation. As explained previously, the underestimation considerablyhigher than the in situ vertical stress. could be due to poor or inadequate sampling by the SEM The majority of the cracksand poresin the samplesappear of the larger aspect ratio cracks in the rock, thereby to have formed early during the rock's history, probably overweighting the low aspect ratio cracks in the distribution during cooling and have sincebeen modified by infillings of spectrum. secondaryminerals. Cracks that are coincident with grain boundariesare abundant in the core. This type of crack is CRACK ORIGIN often produced by thermal stressing[Simmons and Richter, 1976]. Noncoincident grain boundary cracks, thought to Direct observation of micropores and microcracks in the be produced by local strain variations due to the relative Iceland basalts using the SEM supports the idea that the orientation of mineral pairs, are also present.  6656 KOWALL!S ET AL.' MICROCRACKS IN IRDP BASALTS 10-4 10-9 10-2 10-' 1 _ I I i I i i _ Alteration will generally have the effect of shortening long cracks as bridging or partial healing occurs. Samples from _ _ _ _ _ _ _ _ flow tops will be the most porous and generally contain the largest proportions of high aspect ratio pores and cracks, _ _ _ _ _ _ while samplesfrom flow centerswill contain more low aspect ratio cracks. Dike samples contain the fewest cracks and WESTERLY GRANITE have the lowest overall porosity, probably becausethe dikes are intrusive and cooled more slowly under pressure. Thus, they have not been subjected to the same pressure and temperature history that the flow sampleshave. Acknowledgments. We wish to thank N. I. Christensen of the University of Washington for useful discussionand comments and INVERSION for the selectionof the samplesused in the study. Partial financial support was derived from National ScienceFoundation grant EAR ICELAND BASALTS _ 80-08291 and from Gulf Researchand Development Company. _ , I i i I 10-4 10 -:• 10-2 1O- REFERENCES ASPECT RATIO Fig. 6. Range of accuracy for SEM and velocity inversion crack Aki, K., and P. G. Richards, Quantitative Seismology, 557 pp., spectra in Westerly granite [Hadley, 1976; Cheng and Toks6z, W. H. Freeman, San Francisco, Calif., 1980. 1979] and Iceland basalts. Brace, W. F., Some new measurementsof linear compressibility of rocks, J. Geophys. Res., 70, 391-398, 1965. Brace, W. F., E. Silver, K. Hadley, and C. Goetze, Cracks and From the few samples that we observed, it appears that pores: A closer look, Science, 178, 162-163, 1972. Cheng, C. H., and M. N. Toks6z, Inversion of seismicvelocities there exists a progressionof alteration and crack healing, as for the pore aspect ratio spectrum of a rock, J. Geophys. Res., suggestedby Gibson [1979], from the shallower depths where 84, 7533-7543, 1979. alteration rims are common, but very few cracks are bridged Christensen, N. I., and R. H. Wilkens, Seismicproperties, density, or partially healed (as in sample 233.2, Figure 2), to the and composition of Icelandic crust near Reydarfj6rdur, J. deeper flows where alteration rims are not nearly so common Geophys. Res., this issue. Gibson, I. L., Crust of oceanic affinity in Iceland, Nature, 281, (due to the more extensivealteration of the entire rock) 347-351, 1979. but where many cracks are partially or completely healed Hadley, K., Comparison of calculated and observed crack densities (sample 1494.1, Figure 3), This healing has the effect of and seismicvelocities in Westerly granite, J. Geophys. Res., 81, shifting the averagecrack aspectratio to higher values as long 3484-3494, 1976. Meister, R., E. C. Robertson, R. W. Werre, and R. Raspet, narrow cracks are bridged or partially closed. Christensen Elastic moduli of rock glassesunder pressureto 8 kilobars and and Wilkens [this issue]observedan increasein sonicvelocity geophysicalimplications, J. Geophys. Res., 85, 6461-6470, 1980. with depth in the IRDP hole and related it not to a change O'Connell, R. J., and B. Budiansky, Seismic velocities in dry and in porosity but to alteration and low-grade metamorphism of saturated cracked solids, J. Geophys. Res., 79, 5412-5426, 1974. the basalts. Richter, D., and G. Simmons, Microcracks in crustal igneous rocks: Microscopy, in The Earth's Crust, Geophys. Monogr. CONCLUSIONS Ser., vol. 20, edited by J. G. Heacock, pp. 149-180, AGU, Washington, D.C., 1977. Richter, D., G. Simmons, and R. Siegfried, Microcracks, In the Iceland basalts, neither the inversion of velocity micropores, and their petrologic interpretation for 72415 and data nor observations with the SEM provides a complete 15418, Proc. Lunar Sci. Conf. 7th, 1901-1923, 1976. picture of the crack distribution in a sample. High aspect Simmons, G., and D. Richter, Microcracks in rocks, in The Physics and Chemistry of Minerals and Rocks, edited by R. J. G. Strens, ratio cracks that do not close within the range of pressures pp. 105-137, Interscience, New York, 1976. used in velocity data aquisition cannot be distinguishedby Simmons, G., and H. F. Wang, Single Crystal Elastic Constants the inversion theory and many cracks with aspectratios less and Calculated Aggregate Properties: ,4 Handbook, 370 pp., than 0.005 cannot be seen with the SEM. MIT Press,Cambridge,Mass., 1971. Verification of Simmons, G., R. Siegfried, and D. Richter, Characteristics of either technique, using the other, was not possible due to microcracks in lunar samples, Proc. Lunar Sci. Conf. 6th, the small region of overlap in the crack aspect ratio spectra 3227-3254, 1975. produced. However, if both techniques are valid in the Sprunt, E. S., and W. F. Brace, Direct observation of microcavities regions where a crack spectrumis produced, then a complete in crystalline rocks, J. Rock Mech. Mining Sci. Geomech. picture of the crack distribution in a rock is obtained by Abstr., 11, 139-150, 1974. Toks6z, M. N., C. H. Cheng, and A. Timur, Velocities of seismic usingboth techniquestogether. The velocity inversiontheory waves in porous rocks, Geophysics, 41, 621-645, 1976. providesthe low aspectratio portion of the spectrum(below Walsh, J. B., The effect of cracks on the compressibilityof rock, aspectratios of 0.005) and the data from SEM observations J. Geophys. Res., 70, 381-389, 1965. are usedfor the high aspectratio end (>0.005). Wang, H. F., and G. Simmons, Microcracks in crystalline rock from 5.3-km depth in the Michigan Basin, J. Geophys. Res., 83, The nature of the crack aspectratio distribution may vary 5849-5856, 1978. according to the depth of burial or time of burial of the basalt and accordingto the location of the samplein the rock (ReceivedFebruary6, 1981' sequence(i.e., flow top, flow center,or dike). The longeror revised May 11, 1981; deeper the sample is buried, the more alteration will occur. acceptedMay 15, 1981.)