Pdf/83/8/2397/3443198/I0016-7606-83-8-2397.Pdf by Guest on 02 October 2021 2398 GREELEY and HYDE

RONALD GREELEY Space Science Division, Ames Research Center, National Aeronautics and Space Administration, Mqffel Field, California 94035 JACK H. HYDE U.S. Geological Survey and Tacoma Community College, Tacoma, Washington 98465

Lava Tubes of the Cave Basalt,

Mount St Helens, Washington

ABSTRACT tumuli of the Cave Basalt collapsed, probably as a result of withdrawal of supporting lava The Cave Basalt, a high-alumina pahoehoe during drainage of the lava tubes. Raised-rim flow containing numerous lava tubes, originated craters found in many parts of the flow are at the southeast flank of Mount St. Helens, associated with lava tubes and were probably southwestern Washington, and flowed down a formed by collapse of hollow tumuli. stream valley incised in older pyroclastic flow deposits. In situ charcoal samples from two INTRODUCTION localities within lava tubes yield C14 dates of Abundant volcanic structures seen on high- 1,860 + 250 years B.P. and 1,925 + 95 years resolution pictures of the Moon and Mars, and B.P. Detailed survey of 9,125 m of lava tubes, the returned Apollo lunar samples showing correlated with surface geologic mapping, basaltic composition for mare surfaces have yields several geomorphic relations of basalt prompted interest in volcanic landforms as flows. Most of the lava tubes apparently formed analogs to lunar and planetary surface features. between shear planes in laminar lava flow, al- Volcanic geomorphology has been little studied though some tube sections show evidence that recently, particularly in relation to basalt flow the tube roof formed by accretion of spattered surface features. The existence of lunar lava lava in turbulent flow. Partial collapse of tube tubes and channels has previously been pro- interiors reveals: (1) The wall separating the posed (Kuiper and others, 1966; Oberbeck and tube interior from the preflow country rock others, 1969; Greeley, 1971a, 1971d). The ap- may be thinner than 25 cm. (2) Lava flows can plication of analogs to planetary structures re- erode the surface over which they flow. (3) quires a thorough understanding of the ter- Collapse of the tube interior can occur im- restrial counterpart. As part of an investigation mediately after the tube has been drained of molten lava and before the walls cool com- pletely. (4) Lava tubes may represent the thickest part of the lava flow, occupying topographic lows (stream channels). (5) Tubes can be modified extensively through accretion and erosion by later lava flows. Many surface features of basalt flows are directly related to lava tubes. Pressure within the closed lava-tube system caused by outgassing and hydrostatic pressure and overflow of lava from ruptured roof sections (or from channel overflow prior to roof formation) result in formation of a topographic high along many sections of the lava-tube axis. If roof rupture does not occur, tumuli may develop in weak areas of the roof, forming positive features Figure 1. Location map; study area outlines area (may be solid or hollow) 40 to 50 m in diameter covered in Figure 2. Straight-line distance from Port- and several meters high. Most of the hollow land to study area is about 70 km.

Geological Society of America Bulletin, v. 83, p. 2397-2418, 30 figs., August 1972 2397

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 2398 GREELEY AND HYDE

of terrestrial lava tubes and channels, the tubes The first published account of the tubes is an of the Cave Basalt (named here) at Mount St. article in Mazama (1903). Forsyth (1910) Helens in southwest Washington (Fig. 1) were described a lava tube (Ole's Cave) and at- studied in detail. Their excellent preservation tributed its discovery in 1895 to an early makes these tubes particularly well suited for settler, Ole Peterson. However, the other study because they show features that are less major lava tubes remained undiscovered until well exposed in lava tubes of other areas. Unlike the 1950s. Halliday (1963a, 1963b) presented a many basalt flows containing lava tubes, the detailed description of the known tubes, and margins of the Cave Basalt are fairly well de- Pryde (1968) discussed the tubes in a popular fined, and certain relations of the tubes to the article. The only geological descriptions of the configuration of the flow and surface features tubes are by Greeley and Hyde (1970) and can be determined. Hyde and Greeley (1971). Eruptive History and Previous Work Method of Study Although Mount St. Helens has not erupted Prior to field work, photogeologic maps were during this century, it has been active within made of the flow showing the locations of lava- historic time (for a review of historic eruptions tube entrances, surface features, and apparent see Jillson, 1917a; Erdmann and Warren, 1938; flow contacts. In the field, contacts were Holmes, 1955; Folsom, 1970). Published eye- checked and corrected, and additional surface witness accounts indicate that the volcano was features were noted. Following methods de- intermittently active between the mid-1830s scribed elsewhere (Greeley, 1971c), the lava and the mid-1850s. Volcanic ash was erupted in tubes were surveyed in detail (Table 1), and November 1842 and fell as far as 105 km to the where possible, surface features were related to southeast, and lava may have been erupted on the tubes. Rock samples for analysis were the south side of the volcano in 1844. An selected from surface exposures and from within eruption of pumice in about 1800 and an the tube (Fig. 2). eruption of lava on the northwest flank of the volcano sometime between 1800 and the 1830s AREAL GEOLOGY (Lawrence, 1938, 1939, 1941) evidently were The oldest rocks in the map area (Fig. 2) not observed by early explorers or settlers. consist of a series of altered Tertiary volcanic Recent geologic studies (Mullineaux, 1964; and volcanic-clastic rocks that have been in- Hyde, 1970; Hopson, 1971) indicate that the truded by small bodies of andesite, rhyodacite present may simply be the most recent porphyry, and hypabyssal rocks. These rocks quiescent interval between recurring episodes have been studied only in the vicinity of Mer- of volcanic activity. The last few hundred rill Lake (Hyde, 1970) where most of the rocks years have witnessed frequent violent volcan- are tuff-breccia, sandstone, and mudstone, each ism accompanied by hot pyroclastic flows, composed predominantly of rocks and minerals lahars, pumice eruptions, and lava flows. of volcanic origin. Tuff-breccia is predominant. Previous geologic studies of the Cave Basalt Minerals in both matrix and fragments are have been limited to brief observations made by mostly altered to chlorite, carbonate minerals, Verhoogen (1937) and by earlier workers who and clay. Although the ages of the rocks are noted tree wells and lava tubes in the pahoehoe unknown, correlation with similar rocks in basalt (Gibbs, 1854; Elliott, 1897; Diller, 1899; other parts of the Cascade Range suggests that Jillson, 1917b, 1921; Williams, 1922). Erdmann the rocks studied here are late Oligocene and Warren (1938) examined several dam sites (Hopson, 1971, personal commun.). Marble in the vicinity of the volcano and drew at- Mountain, about 10 km southeast of Mount tention to possible volcanic hazards. Lawrence St. Helens, consists mostly of nearly uniform (1938, 1939, 1941, 1954) and Carithers (1946) olivine basalt flows. The basalt flows typically described recent pyroclastic deposits on the are 1 to 3 m thick and consist of layers of north side of the volcano. Mullineaux and clinkery scoria alternating with light-gray to Crandell (1962), Mullineaux (1964), Hyde dark-gray basalt. The basalt is massive to platy (1970), and Hopson (1971) have discussed var- and dense to porphyritic with large laths of ious geologic aspects of the volcanic deposits. plagioclase and crystals of olivine. One sample The lava tubes in the Cave Basalt were ap- was selected for chemical analysis (sample 6, parently unrecognized by earlier observers. Table 2); it is distinguished from other basalts

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 GEOLOGICAL MAP OF SWIFT CREEK AREA MOUNT ST. HELENS. WASHINGTON R. Greeley and J. Hyde, 1971

EXPLANATION

|::::::'ri";:*l Undivided surficial deposits; alluvium, lahars, pyroclastic flows, E&w?.ja and glacial drift.

Quaternary ''Q\'/\ Undivided Mount St. Helens lavas; pahoehoe, aa, and block basalts, andesites and dacites. MN Qcb j Cave Basalt; olivine basalt, primarily pahoehoe, some aa.

21%° Quaternary (?) Marble Mountain Basalt; olivine basalt and andesite.

Tertiary TV I Volcanic and volcanic-clastic deposits, with some intrusive rocks. 1958 ^_^ Contact, dashed where covered, dotted where approximate or uncertain.

O Sample locations. 122° 22' 36" W K Dip and strike of beds. 46° 09' 45" N R-«.R.5E Lava tube Tumuli

0 1 2 3 I . 1 1 , Miles 0 1 2 30 4 I.I 1 1 I Kilometers

Base from USGS Cougar and R.4E R.5E 122° 07' 13" W Mt. St. Helens 15' quadrangle and 46° 02' 32" N U.S. Forest Service maps. Figure 2. Geological map of the study area.

GREELEY AND HYDE, FIGURE 2 Geological Society of America Bulletin, v. 83, no. 8

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 LAVA TUBES OF THE CAVE BASALT, MT. ST. HELENS, WASH. 2399

TABLE 1. LOCATION AND PHYSICAL CHARACTERISTICS OF LAVA TUBES IN THE CAVE BASALT

Maximum Elevation Length* Name Location width Average slope (ft) figr (m) (m)

Ape Cave8 NEHSWWEH sec. 8, T. 8 N., R. 5 E. 2,115 3,400 11.6 12.2 3.3°, 1 m/17.2 m Barney's Cave NWW>sSW>< sec. 17, T . 7 N., R. 5 E. 1,670 56 2.7 3.6 2.0°, 1 m/27.8 m Bat Cavef sec. 19, T. 7 N., R. 5 E. 1,160 330 3.7 16.2 Beaver Cave SE«E!» sec. 18, T. 7 N., R. 5 E. 1,560 477 9.1 15.2 3.0°, 1 m/19.6 m Dollar-Dime Cave NlfeSEljNWk sec. 19, T . 7 N., R. 5 E. 1,290 -- -- — — -- Flow Cave EyWWVfc; sec. 32, T. 8 N., R. 5 E. 2,835 99 2.4 4.6 3.2°, 1 m/17.4 m Gremlin Cave1 NtWIEWIEl< sec. 31, T . 7 N., R. 5 E. 3,000 ------— Lake Cave SiSWtSDt sec. 8, T. 7 N., R . 5 E. 1,845 1,248 15.5 9.1 2.6°, 1 m/22.2 m Little People Cave W>sNEIiSW!» sec. 32, T. 8 N., R. 5 E. 2,835 143 4.6 7.9 1.3°, 1 m/47.7 m Little Red River Cave WiNE>tSU>i sec. 32, T. 8 N., R. 5 E. 2,820 1,032 9.1 11.9 4.5°, 1 ra/11.7 m Die's Cave1 SW«Wy«\ sec. 20, T. 7 N., R. 5 E. 1,370 1,592 7.6 13.1 2.1°, 1 m/27.4 m Powerline Cave N>5 sec. 30, T. 7 N. , R. 5 E. 1,140 '- Prince Albert Cave sec. 19, T. 7 N., R 5 E. 1,180 480 6.1 15.0 - Spider Cave sec. 32, T. 8 N., R. 5 E. 2,840 271 4.6 11.4 Elevations are estimates determined from U.S. Geological Survey Cougar and Mount St. Helens 15-ndn quadrangles. *For lava tubes with multiple passages, length represents main passage only. Dimensions from Halliday (1963a). Location given for main entrance. in the region on a plot of total alkalis and silica entrance to Lake Cave. Other common surface (Fig. 3). features include ropy or corded textures, pres- The Cave Basalt, here named and defined, is sure ridges, and tumuli that are often cracked a pahoehoe basalt flow that originated at the or collapsed. The term tumulus is used here for southwest flank of Mount St. Helens—its type any dome-shaped, circular or semicircular, locality—and flowed south about 11 km down solid or hollow surface feature. Most of the a stream valley cut into pyroclastic flow known hollow tumuli of the Cave Basalt are deposits of late Quaternary age. The upper part collapsed, leaving raised-rim craters up to 50 m of the lava flow (Figs. 2, 4) is covered by in diameter. younger lava and pyroclastic flows and is Excellent exposures on the bluff overlooking poorly exposed. The Cave Basalt may be the Lewis River valley and in segments of the present as high as 1,465 m elevation on the lava tubes indicate that the basalt flow filled a volcano and can be traced to the Lewis River stream channel incised in a relatively flat- valley where it terminates on the north bank floored valley. At station 34 (Fig. 5) in Lake of the river. The basalt is light gray to dark Cave, breakdown of the east wall of the lava gray, vesicular to massive, containing pheno- tube reveals the edge of the stream channel crysts of olivine and plagioclase. In thin section about 3 m above the tube floor. Erosion by the plagioclase is usually euhedral, slightly running water at the base of the lava flow zoned, ranges from about An70 to An40, and makes it possible to crawl about 30 m under the averages 1 to 2 mm long. The olivine pheno- relatively flat floor of the lava flow where the crysts average }/j to 1 mm and generally are undersides of tree molds are exposed. The basalt broken along curved fractures into several flow was temporarily ponded on the bluffs fragments. The vesicles are spherical in con- overlooking Lewis River; apparently a dam trast to the more irregular shaped vesicles of aa was created in the relatively narrow channel basalt, range in size from about 0.5 to 8 cm, and incised in the pyroclastic flow deposits. Craters make up 1 to 20 percent of the rock. The three as large as 50 m in diameter in the ponded lava chemical analyses (samples 1, 2, 7, Table 2) are were created by collapse of tumuli that ap- remarkably uniform and all fall within the parently were inflated by hydrostatic pressure high-alumina basalt field of Kuno (1968). The resulting from lava in the tubes, as discussed basalt flow surface is generally hummocky or below. The age of the basalt flow is probably flat over broad areas and is commonly broken close to 1,900 years B.P. Charcoal from roots at into large tilted slabs. Vertical and horizontal the base of a tree mold exposed in the side tree molds are common near the flow margins passage at station 34 in Lake Cave was dated as and are particularly well displayed near the 1,860 ± 250 years B.P. (W-2270). Charcoal

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 2400 GREELEY AND HYDE

similarly exposed at the base of the lava flow covered interval about 5 km wide; however, in Ape Cave was dated as 1,925 + 95 years lava of the Cave Basalt is present at the drain- B.P. (GX1673). We found no evidence for age divide and in the upper part of the Kalama intervening periods of surface exposure and River drainage basin, so the two flows may be weathering; the flow was evidently extruded connected. One sample of basalt collected from over a short time interval and may have acted the Kalama valley (sample 3, Table 2) is chemi- as a single cooling unit. cally very similar to samples of Cave Basalt. A basalt flow exposed in the Kalama River Rocks mapped as undivided lava of Mount valley and in the vicinity of Merrill Lake (Fig. St. Helens include basalts, andesites, and dacites 2) has the same characteristics (Hyde, 1970) as that crop out in the Swift Creek and Kalama the Cave Basalt and probably was extruded at River valleys. A block lava flow at the head of the same time. The two flows are separated by a the Kalama River consists at the surface mostlv

TABLE 2. CHEMICAL ANALYSES AND MOLECULAR NORMS

Chemical analyses (weight percent) Sample 1 2 3 4 5 6 7 8 9 10

S102 50.13 50.28 49.59 62.46 54.09 49.22 50.15 63.16 54.20 49.3 Ti02 1.65 1.46 1.60 0.92 1.51 1.78 1.02 0.54 1.31 1.6

A1203 16.73 17.19 17.47 16.82 16.73 17.98 17.81 18.22 17.17 15.6

Fe203 2.53 1.72 3.32 1.40 2.84 6.16 1.16 1.36 3.48 3.5 FeO 7.94 8.48 7.46 4.28 6.29 3.37 8.82 3.33 5.49 7.8 CaO 9.42 9.72 9.69 5.16 7.76 7.64 10.13 5.24 7.92 10.3 MgO 7.10 6.98 6.62 2.12 5.39 7.18 6.96 2.30 4.36 6.5 MnO 0.15 0.14 0.15 0.08 0.13 0.15 0.18 — 0.15 0.2

Na20 3.31 3.38 3.40 4.72 3.92 3.45 3.28 4.06 3.67 2.7

K20 0.49 0.50 0.52 1.52 1.06 1.14 0.41 1.16 1.11 0.5

H20+ 0.82 0.56 0.27 0.60 0.59 0.67 0.40 0.86 1.8

H20- 0.33 0.19 0.13 0.40 0.22 1.00 0.08 0.10 — --

P205 0.24 0.21 0.16 0.24 0.32 0.38 0.15 0.14 0.28 0.3

Total 100.84 100.81 100.38 100.72 100.85 100.12 100.15 100.01 100.00 100.1

Molecular norms

Q ------13.05 2.58 -- 5.7 c — -- 0.43 -- -- — — Or 2.91 2.95 3.06 9.01 6.26 6.81 6,7 Ab 29.80 30.25 30.32 42.51 35.19 31.30 30.9 An 29.43 30.16 29.61 20.28 24.92 30.52 27.2 Wo 6.35 6.62 — 1.53 4.60 1.19 4.2 En 13.83 9.98 16.70 5.88 14.88 20.27 10.9 Fs 5.98 5.24 4.53 4.52 5.87 - 5.3 Fo 4.38 6.93 7.41 — -- 0.79 - Fa 1.89 3.64 2.01 — - — - Mt 2.66 1.80 3.45 1.47 2.97 4.51 5.1 Hm ------1.34 — n 2.31 2.03 2.22 1.29 2.11 2.51 2.4 Ap 0.51 0.44 0.34 0.51 0.67 0.81 0.7

Sample description and location

*1 Cave (?) Basalt, Swift Creek, SWijNEJs sec. 30, T. 8 N., R. 5 E.; this study. *2 Cave Basalt, Lewis River, NWJjSWij sec. 25, T. 7 N., R. 5 E.; this study. *3 Undivided lava of Mount St. Helens, Kalama River, SwWt sec. 33, T. 8 N., R. 4 E.; this study. *4 Undivided lava of Mount St. Helens, Swift Creek, NWjSBs sec. 25, T. 8 N., R. 4 E.; this study. *5 Undivided lava of Mount St. Helens, Swift Creek, NEydfe sec. 32, T. 8 N., R. 5 E.; this study. *6 Basalt of Marble Mountain, Swift Creek, NEWIBt sec. 16, T. 7 N., R. 5 E.; this study. 7 OHvlne basalt (Cave Basalt), pahoehoe flow, Lewis River; Verhoogen (1937, p. 293). 8 Pyroxene andesite, summit, Mount St. Helens; Verhoogen (1937, p. 293). 9 Average andesite; Nockolds (1954, p. 1019). 10 Average Picture Gorge Basalt; Waters (1961, p. 595). 'Analysis by shiro Inad of the Japan Analytical Chemistry Research Institute.

of relatively smooth polyhedral blocks ranging in size from 0.3 m to 2 m on a side. The rocks DA have a dark glassy luster, often show con- choidal fracture patterns, and are aphanitic. The interior of the flow is well exposed in a road cut on the south edge of the flow where 1* only brecciated material is visible. The flow o has abrupt margins and is about 10 m thick. The block flow has a distinctive chemical composition (sample 4, Table 2; Fig. 3), but its age relation to the Cave Basalt is not clear. Aa lava comprises the eastern lobes of the 45 50 60 undivided lavas of Mount St. Helens except SiO2 (weight percent) for andesite belonging to an ancestral St. Helens sequence, just east of Swift Creek O Cave Basalt!?), sample 1 (Hopson, 1970, personal commun.). The aa O Cave Basalt, sample 2 flows are characterized by jagged, clinkery sur-

Figure 4. Oblique aerial overview of the Cave Swift Reservoir in foreground (courtesy U.S. Forest Basalt; Mount St. Helens in background; spillway of Service).

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 2402 GREELEY AND HYDE

carbon age of 1,740 ± 250 years (W-2527). tain. These are the only tubes known west of Pyroclastic layer W, erupted from Mount St. the kipuka. Helens about 450 years ago (Crandell, 1969), Power Line Cave, Beaver Cave, and Barney's lies on the surface of the flow. Cave (Figs. 11, 12) are lava tubes apparently Extensive deposits of volcanic ash, pyroclas- not associated with any particular system. A tic flows, and lahars (Mullineaux and Crandell, collapse pit, a short lava-tube segment, and 1962; Hyde, 1970; Hopson, 1971) resulting several small tubes (Halliday, 1971, personal from explosive eruptions are present in the commun.) in the northern part of the flow also map area, but discussion of these deposits is seem unrelated to any system. beyond the scope of this paper. Formation of Lava Tubes LAVA TUBES Lava-tube formation is apparently controlled primarily by lava-flow viscosity, which Lava-Tube Locations in turn is related to the chemistry, dissolved Figures 2 and 6 show the location of the lava gasses, temperature, and flow velocity of the tubes within the Cave Basalt; Table 1 lists their lava. In the Cave Basalt, lava tubes occur only geographic coordinates and physical character- in pahoehoe, the fluid variety of basalt. At istics. The principal lava-tube system in the least two modes of tube formation are displayed flow is the Mount St. Helens system (Halliday, in the Cave Basalt, and both may occur within 1963a, p. 97) which exceeds 8,333 m in length a single tube in different localities. The first and is composed of Little Red River Cave, Ape mode of formation is by constructional ac- Cave, Lake Cave, and Ole's Cave. Although cretion of the lava-tube roof. Observations of the system begins in the constricted northern active lava-tube formation in Hawaii (Greeley, part of the flow, it probably originated up- 1971b) show that rapidly flowing lava on steep slope, above Little Red River Cave. The sys- slopes is turbulent and follows narrow chan- tem trends southward along the east flow nels. Splashing and spattering lava accretes margin and apparently terminates in collapsed along the channel edges to form arched levees tumuli east of Green Mountain kipuka. Ape that eventually merge over the channel, form- Cave, with a passage length of 3,400 m, is the ing an arched roof. The roof may remain after longest known uncollapsed lava-tube segment; drainage of molten lava at the close of the however, longer, partly collapsed tubes are eruptive stage. Parts of Little Red River Cave known in many areas (Hatheway and Herring, probably developed by such accretion. Stations 1970). Hopeless Cave, several dozen meters 1 and 4 (Fig. 5) mark very steep lava-tube north and slightly west of the main entrance to sections in which partial roof collapse reveals Ape Cave, may have drained into Ape Cave in agglutinated clots of lava probably formed by the late stage (Halliday, 1971, personal accretion. This structure contrasts with the commun.). It may or may not have been part more regular layered lava found on less steep of the main system. A small group of lava tube segments described below. tubes in the northern part of the flow (Gremlin The second mode of formation occurs in low Cave, Spider Cave, Flow Cave, and Little velocity, laminar lava flows such as those People Cave; Figs. 7, 8) may be related to the described by Oilier and Brown (1965) and Mount St. Helens system. Spider Cave and Hatheway (1971). Shear planes are believed to Flow Cave align with the general trend of form within the flow (Fig. 13), and as the main Little Red River Cave and, as discussed below, body of the flow cools, active flow is restricted may have fed lava to the main system during to the hottest, thickest part. At this stage the the final eruptive stages of the Cave Basalt. exact position of 'the mobile cylinder is not Three lava-tube segments are west of Green fixed, and the primitive lava tube can shift Mountain: Dollar and Dime Cave, Prince about (that is, surges of lava from the vent can Albert Cave, and Bat Cave (Figs. 9, 10). Al- cause the tube to migrate from one part of the though these tubes were not examined, discus- flow to another). As the lava progressively sions with local spelunkers (C. V. Larson, 1969, cools, the structure becomes more firmly fixed; personal commun.) and analyses of diagrams eventually the bulk of the flow solidifies, and photographs lead us to believe that these forming layered lava, and the still-molten tubes are part of a system extending along the material drains from the most active part of the eastern contact of the flow with Green Moun- flow and leaves a void or lava tube.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 &ana MOOT [^gnaj Coltac* Mocks 2700 (§) Station numbers Vertical exaggaration «2 Slabs spelled from uppir walls and roof; lining up to 15-28 cm thick UPPER APE CAVE SECTION Country rock, both walls Skylight , Upper tube Crocs Mctfon AA' Country rack exposed intermittently In troth went

Tubs incised in hillsids behind lining ©

Large breakdown btocki Groovad lava in fused in floor flow Wall forme d «««•(l» an llnd mural* LOWER APE CAVE SECTION k in lava Traamold gunan °" " Tuba lining with viacoelattic ructure, country rock in west wall Tuba lining 20 cm thick, walls. poaaiUa ajrga contact Cupota, aurfaca expression 2100 as dome, 20 m in diameter Slabs of lining spelled fled breccie, from walls and transported wast wall. Lining by lava 20 cm thick.

LAKE CAVE SECTION 1800

Collapse blocks fused to walls OLE'S CAVE SECTION Cupola Surface collapse Entrance No. 2 (possible vent)

WWWJJMH/* >>t»S, Lateral breakdown exposures, red-brown clinkers behind tuba wall.

1 profile of the Mount St. Helens kva-tube system. GREELEY AND HYDE, FIGURE % Geological Society of America Bulletin, v. 83, no. 8

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 CAVE BASALT FLOW - SURFACE DETAIL R. Greeley and J. Hyde

Figure 6. Topographic map of the Cave Basalt.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 2404 GREELEY AND HYDE

we believe parts of the lava tube are situated in the former stream channel of the valley. As such, the configuration of the channel would (to some degree) determine the shape and ground plan of the lava tube. Interaction of molten lava with stream water, snow, or ground water may lead to phreatic explosions within the lava flow. One area of Ape Cave shows evidence (Hyde and Greeley, 1971) that a phreatic explosion ruptured the wall of the tube and introduced steam charged with particles of country rock (clay), which fused on the tube walls (Fig. 15). Although such events are probably rare, they can significantly deform lava tubes. As the lava flow advances and spreads later- Figure 7. Diagram of Spider ally, the main lava tube supplies molten lava Cave, plan view (WJ^NW^SWJi to the flow front through a series of distributary sec. 32, T. 8 N., R. 5 E., elev. 2,840 feeder tubes. During drainage of the main tubes, ft, Skamania County, Washington, feeder tubes are often plugged or only partly from survey by D. Packard, L. Mills, drained. Lava tubes may bifurcate to form two and J. Hilton). distinct networks, or, as with river channels, Lava tubes seem to develop along the most may form cutoff branches (Fig. 9). rapidly flowing part of the active lava flow. In the later stages of tube formation, prob- Figure 6 shows that tubes are not necessarily in ably during drainage, a layer of lava is accreted the center of the flow; rather, they are common along the walls and ceiling to form a lining along the flow margins. Similar relations have (Fig. 16). Lava-tube linings range from less been noted in lava tubes in California (Greeley than 1 cm to more than I m thick and are and Baer, 1971). Configuration of the preflow generally somewhat more vesicular than the surface probably controls the overall trend of main body of the lava flow. Vesicularity within lava tubes. The Cave Basalt occupies a former the lining may increase toward the tube, stream valley, as reconstructed in Figure 14. possibly representing outgassing into the tube. From evidence discussed in the next section, During or after drainage, parts of the lining

Elevation, Tt 2850

Crawlway

No vertical exaggeration Figure 8. Cross section and plan view of Flow Cave ft, Skamania County, Washington, from survey by R. sec. 32, T. 8 N., R. 5 E., elev. 2,835 Greeley and M. Lovas).

may develop a glassy surface. Glaze that is deformed (Fig. 17) must have been formed while the lining was still plastic and subject to movement. Formation of glaze is not well under- stood ; however, it appears to develop as a thin coating over the outer vesicular zone of the lining (Greeley, 1969), possibly as a result of rapid cooling by air currents in the lava tube. As lava drains from the tube, a temporary Figure 10. Diagram of Bat Cave, plan view NWJiSWJi sec. 17, T. 7 N., R. 5 E., elev. 1,160 ft, Skamania County, Washington, from HaUiday, 1963a). cessation of drainage can result in horizontal flow lines along both walls. Successive halts produce a series of flow lines (Fig. 18) that can be traced and correlated for hundreds of meters along the tube. Incomplete drainage of low- lying parts of the tube or deformation of the lava-tube roof may close tube passages with structures termed lava seals by Halliday (1963a, p. 124). The upper ends of Ape Cave and Little Red River Cave are blocked with lava seals (Fig. 5). Relation of Lava Tubes to Preflow Geology Many wall sections of the Mount St. Helens lava-tube system have collapsed to expose preflow country rock (Figs. 5, 19). In most exposures, country rock deposits (pyroclastic flows) are baked a brilliant red. These exposures (rare in most other areas) provide a unique opportunity to study relations of lava tubes and flow to preflow topography. For example, station 31 in Lake Cave shows that there may be no more than 2 cm separating the lava-tube interior from the outer edge of the flow. Figures 2 and 5 show that many collapses of tube walls occur where the tube is close to the flow edge. Figure 14 shows composite cross sections of parts of the system. From the exposures available, we conclude that parts of the lava tube are situated in a prelava-flow stream channel. This channel is about as deep and wide as the present stream channel of Panamake Creek, a small stream a few miles west of the Cave Basalt but unaffected by Holocene volcanic eruptions. Sections of tube along the presumed stream channel assume a skull-shaped cross section; undercutting by the flow may form asymmetric sections. Some sections in the middle of Figure 9. Diagram of Prince Albert Cave, plan the flow on gentle gradients have horizontally view (NKNWMSWi^ sec. 19, T. 7 N., R. 5 E., elev. oval sections which may be due, in part, to 1,180 ft, Skamania County, Washington^/rom survey by lateral erosion. Tube configurations in meander C. Larson and J. Larson). bends are often asymmetric, forming "cut-

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 2406 GREELEY AND HYDE

Tree mold in ceiling

•g 1500 -

^^^^- 1400I— South North -11400 Longitudinal cross section Vertical exaggeration x 2 Figure 11. Cross section and plan view of Beaver ft, Skamania County, Washington, from survey by R. sec. 18, T. 7 N., R. 5 E., elev. 1,560 Greeley and M. Lovas).

banks" similar to stream beds. Further evi- parently had two or more lava surges or flows dence of erosion by lava is found in the form that flowed both through an existing tube and of large inclusions of country rock in the lava perhaps over the surface of the tube. These flow at station 15, Ape Cave. later flows apparently remelted the existing Exposures in Little Red River Cave and Ape roof to extend the tube vertically, forming a Cave (Figs. 5, 14B, 19) show country rock vertically elongate cross section. In some areas overhanging the lava tube. It is unlikely that of Ape Cave (station 10) only parts of the roof this was the original configuration of the were eroded by surface flows, forming distinct topography in an area of easily eroded uncon- upper-level lava tubes. Flow patterns in the solidated deposits; these overhangs probably floor of the upper-level tubes show that when Elevation, ft. r represent erosion by the lava flow. Sections of Ol Little Red River Cave, for example, undercut 1670 r- [ parts of the eastern, preflow hillside (Fig. 14). Thus, to some degree the position of the tube is determined by the erosive capability of the lava flow. Subsequent Lava Flows Lava tubes often serve as conduits for later lava flows and subsequently are usually altered. Alteration may be in the form of filling or partially filling the tube (Fig. 20), remelting the lava-tube roof to form vertically elongate tubes, reshaping or eroding the tube walls, or any combination. Figures 14 and 21 show several tube sections Figure 12. Cross section and plan view of Barney's that appear to have resulted from subsequent Cave (NWJiNEMSWJi sec. 17, T. 7 N., R. 5 E., elev. flow alteration. In cross section, the tube near 1,670 ft, Skamania County, Washington, from survey the main entrance of Ape Cave (Fig. 18) ap- by R. Greeley and M. Lovas).

------

Figure 13. Diagrams illustrating lava-tube for- transverse cross section (normal to axis, lower diagram). mation associated with shear planes developed in The tube cross section in stage 4 is shown with an laminar lava flow. Each stage is shown in longitudinal hourglass outline to emphasize that noncircular tubes profile (parallel to flow axis, top diagram) and in may form within a single flow. the system drained, lava flowed from the upper within a single massive flow. As cooling pro- level into the lower, main tube through eroded gressed, the conduits may have merged to roof sections. form single, elongate tubes, or may have de- An alternative explanation for multiple-level veloped as individual tubes stacked vertically. tubes and vertically elongate cross sections is In some areas, the conduits may have merged that individual conduits of lava developed and separated alternately along the course.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 2408 GREELEY AND HYDE

Drainage of subsequent lava flows from an older lava tube may result in the accretion of additional linings. For example, Mud Cave at the National Reactor Test Site, Snake River Plain, Idaho, is a lava tube that shows at least three individual linings along much of its length. In theory, it should be possible to determine the number of flows that have occurred by counting the number of linings in a wall section. However, the linings often merge to form one lining, are remelted, or are com- pletely destroyed. Small subsequent flows are often found along the floors of lava tubes. These flows may be either pahoehoe or aa and may display flow patterns. As with their larger counterparts, small flows may develop channels, levees (Fig. 22), or small lava tubes, resulting in "tube-m- tube" structures (Fig. 23). Interior flows may erode the walls of the larger tube, tearing away parts of the lining and rafting the blocks down- tube. Blocks of rafted material are fused to the floor flow in several areas (for example, near the entrance to Little Red River Cave). In

PRE FLOW EXPOSURE FUSED CLAY

Figure 14. Transverse cross sections, Mount St. Helens lava-tube system, showing relations of lava tube to lava flow and preflow country rock. In these sections, the lava tube is interpreted to occupy a former stream channel. In some instances, subsequent flows intersect and merge with existing lava tubes at some distance from the vent (Greeley, 1971b). For example, the western passage of Lake Cave (Halliday, 1963a, p. 93-96) appears to have been a small (2 m wide X 1 m high, maximum) lava tube near the surface that drained into the main tube. A similar formation probably occurred at the upper end of Little Red River Cave. The lava is distinctly red in the upper level, in contrast to the black lava of the lower level, and matches in color and appearance the lava of Flow Cave, a few hundred meters up- flow from Little Red River Cave. Flow Cave is a smaller lava tube formed near the surface, 100 meters similar to the west branch of Lake Cave. We Figure 15. Section in Ape Cave; plan view showing believe Flow Cave drained into and eventually the probable source of fused clay material and site of sealed the upper end of Little Red River Cave. phreatic explosion.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 Figure 16. Wall section of Lake Cave a few hundred ruptured and slid to the floor; molten lava behind the meters from the entrance. Lining (about 20 cm thick) lining spilled over the lining and into the tube interior.

Figure 17. Deformed glassy wall glaze on lava-tube tion indicates flowage of the lining before it had corn- lining near station 4, Little Red River Cave. Deforma- pletely solidified.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 Figure 18. Lava-tube interior near main entrance to and multiple flow lines representing successive stages of Ape Cave showing complex cross section, lateral gutters, drainage. Tube is about 13 m high.

Figure 19. Near station 1, Little Red River Cave. lava flow. Dark band along contact is baked zone about A. Contact of lava on right side and country rock on 20 cm wide. B. Uncollapsed continuation of the tube. left side, indicating undercutting and erosion by the Arrow at bottom of picture marks a hardhat for scale.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 Figure 20. Spider Cave in which a subsequent floor graph courtesy of C. Larson. flow of aa has partially filled the tube interior. Photo-

Figure 21. Transverse cross sections of lava tubes station 28, down-tube view: E. Little Red River Cave showing "cutbank" configurations and multiple-level station 47, down-tube view: F. Ape Cave station 26, up- development. Stippled area represents lava containing tube view: G. Ape Cave station 32, up-tube view: H. the tube: A. Lake Cave station 15, up-tube view: B. Ape Cave station 37, up-tube view: I. Ape Cave station Lake Cave station 24, up-tube view: C. Lake Cave 1, down-tube view: J. Ape Cave station 48, up-tube station 39, down-tube view: D. Little Red River Cave view.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 2412 GREELEY AND HYDE

some sections, large clots of lava are fused on the walls (Fig. 24). Lava-Tube Deformation Viscoelastic deformation of the walls, lining, or roof commonly occurs during and immediately following drainage of lava from a tube. Figure 5 identifies some examples of wall deformation. In these areas (Figs. 16, 25) the lining tore away from the wall and slid toward the floor. Molten lava behind the lining can flow through the rupture and dribble inside the tube. Station 17 (Ape Cave) marks an exposure of country rock in which the lining and wall collapsed, allowing country rock to slump into the tube. Molten lava then dripped on the country rock and cooled, forming small blobs of glassy basalt. Similar occurrences are found along sections of Lake Cave. Several collapsed Figure 22. Floor flow illustrating development of sections of the lining of the Mount St. Helens prominent levees along a deep, sinuous channel. Photo lava-tube system expose brick-red lava clinkers by W. Halliday. (autobrecciated lava) between the lining and the have occurred by vesiculation into possible massive basalt of the flow. Differential move- voids and low-pressure areas between the lining ment of the lining and massive basalt may and the body of the flow, similar to the process produce a shear zone along which the auto- described by Parsons (1969) for lava flows in brecciated lava forms. Brecciation may also general.

Figure 23. Tube-in-tube structure, south of and channels in floor flows is relatively common in lava entrance 2, Ole's Cave. Formation of small lava tubes tubes.

Viscoelastic deformation, shown by inward draping of layered lava toward the entrance to Ape Cave, appears to have resulted from roof collapse during or immediately following drainage. Although rare in the lava tubes of the Cave Basalt, this type of roof collapse is fairly common in other areas (Greeley, 1971c). Oc- casionally the roof may sag slightly and fracture along tens of meters of the tube axis. Although medial fractures usually develop along the axis of the sag, they may also form in tubes without roof sag. Other fractures often develop in the lining in addition to the medial fracture to form polygonal fracture patterns. Some parts of lava tubes seem to have been deformed by pressure on the upper, viscoelastic walls and roof. Hydrostatic pressure within the closed tube system and outgassing into the tube interior probably were sufficient to push ceiling sections into bulbous chambers, or cupolas (Fig. 5). In some places, pressure apparently was great enough to rupture the roof, as described below. Lava-tube collapse after cooling begins with Figure 24. Station 20 in Ape Cave. Unusual lava- spalhng of the tube lining. This is followed by tube formation in which a ball of lava wedged between spalling and collapse of basalt blocks from the and fused to the tube walls during drainage of molten roof, eventually forming a small opening material from the tube.

Figure 25. Tube wall near station 2, Little Red collapsed during the final stages of drainage and ex- River Cave, illustrating extremely thin (less than 2 cm) posed country rock. Glove indicates scale. wall separating tube from country rock; tube wall

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 2414 GREELEY AND HYDE

(termed a skylight) to the surface. Some parts Deformation of plastic semimolten lava-tube of lava tubes are marked by extensive roof lining can produce wall drapery along several collapse (not necessarily breaking through to meters of tube wall. Overhanging sheets of lava the surface, however) that may block the in- (Fig. 27), less than 1 or 2 cm thick, produce terior passage. Percolation of surface water into ribbon stalactites (Halliday, 1963a). Unusual tubes and growth of vegetation in the roof stalactites in Flow Cave are deformed down- contribute to weathering and collapse of lava tube in a cross-bedded pattern, possibly de- tubes. Small roots from surface vegetation are formed by strong air or gas currents flowing visible in Ape Cave. The flow of water through through the tube during the later stages of tubes of the Cave Basalt is often voluminous, drainage. particularly during the spring runoff. Little Red River Cave was named from a small red SURFACE FEATURES OF stream that runs down the tube for about two- LAVA TUBES thirds of its length. Water stained with red One objective of our study was to determine clay from the baked country rock enters the the relation of lava tubes to the surface tube in a series of springs near station 5. The topography of the lava flow. Many surface stream flows all year and is apparently fed features on the Cave Basalt can be correlated from surface water and ground water from the directly with lava tubes. Lava tubes often form Mount St. Helens snow melt. subtle ridges along their axes (Greeley, 1971c) Seasonal streams flow in most of the other because they usually occupy the thickest part Cave Basalt lava tubes. Sand and other of the flow. Distributary tubes and roof detritus are washed into the tubes or eroded ruptures also result in accretion of lava along from country rock exposures and transported the main tube axis. Figure 14 illustrates cross by the streams. At the tube ends (Little Red sections made transverse to the axis of Lake River Cave, Lake Cave) and in depressions Cave and Ape Cave showing their topographic along tubes, stream water collects in small expression. On flows known to contain lava ponds and deposits sediments. Changes in hy- tubes, ridges parallel to the direction of flow drologic conditions within the tube can result may indicate subsurface tubes. in rejuvenation of the stream, with subsequent Occasionally, pressure is adequate in an down-cutting through the deposits. Sections active lava tube to force lava through roof near station 30 in Ape Cave show eroded ter- cracks, or perhaps to rupture the roof, and races with more than a meter of graded silt, spill lava on the surface as a small flow. Some- sand, and gravel stream deposits. times a small hornito may develop over the Continued erosion and weathering lead to roof. Entrance 4 to Ole's Cave and the section complete collapse of the tube roof to form a near station 32 of Lake Cave mark areas where sinuous trench. Because of their young age and small lava flows erupted from the tube. Many relatively thick roofs, tubes in the Cave Basalt apparently anomalous aa flows found as scat- have not yet progressed beyond the skylight tered patches in the middle of pahoehoe may stage. have flowed from lava tubes, although detailed mapping of the patches showing their possible Minor Features relations to tubes has not been done. During final drainage of molten lava from In some instances, pressure within the tube the tube, several varieties of lava cave forma- was not sufficient to rupture the roof. Instead, tions may develop. Lava dripping from the the roof was inflated to form a cupola. The ceiling can produce lava stalactites and stalag- surface expressions of cupolas are often small mites; both structures can be tens of centi- tumuli. Station 19, Ape Cave, marks a cupola meters long, although the normal length and corresponding tumulus. Most hollow seldom exceeds 8 or 9 cm. Excellent examples tumuli apparently collapsed almost immedi- of stalactites and stalagmites are in the Cave ately following their formation, indicated by Basalt lava tubes. Figure 26 illustrates in- plastic deformation of the crust. The resulting cipient lava stalagmites that formed as the structures are craters as large as 50 m in diame- floor flow moved downslope, as indicated by the ter with raised rims. Figure 28 shows the loca- linear array of glassy blobs on the flow surface. tion of craters and tumuli in relation to part of Cessation of flow permitted the dripping lava Lake Cave. Two prominent collapse craters to accrete in one place and to form the stalag- (Fig. 29) mark the end of Ole's Cave. We mite. believe these craters resulted from temporary

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 Figure 26. Section of floor in Spider Cave. A floor a chain of glassy blobs (A); cessation of flow permitted flow apparently was moving down-tube while molten a small kva stakgmite to accrete (B). Photo courtesy of lava dripped from a single source on the ceiling, forming C. Larson.

Figure 27. Ribbon stakctite near station 2, Little drained from behind a ruptured part of the tube lining. Red River Cave, formed by a thin sheet of kva that

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 2416 GREELEY AND HYDE

1 Flow direction

Entrance

100 200 meters Figure 28. Diagram of part of Lake Cave, showing tumuli; hachures indicate degree of collapse. associated collapsed, partly collapsed, and uncollapsed ponding of the Cave Basalt before the flow tumuli (Fig. 30) and collapse craters have spilled into the Lewis River valley. During small (up to 4 m wide) lava tubes leading down- ponding, lava continued to flow into the pond slope away from the rim. These probably de- beneath the crust under hydrostatic pressure veloped during collapse when parts of the rim from the lava tube (similar to the way in which ruptured and molten lava flowed on the surface. lava tubes and collapsed lava ponds formed in Numerous pressure ridges occur on the Cave central Oregon [Greeley, 1971c]). Release of Basalt. Although a rigorous analysis of the size, the flow into the valley relieved the pressure; location, and orientation of pressure ridges was as the lava drained from beneath the dome not made, there is no apparent relation of the crust, the structure collapsed to form a crater. pressure ridges to the tubes. A large pressure ridge transverse to flow direction is near the south end of the flow (Fig. SUMMARY AND CONCLUSIONS 6). This structure, about 450 m long, 35 m The Cave Basalt, a high-alumina pahoehoe wide, and as high as 4 m, is believed to have flow, originated at the southwest flank of formed as a result of downslope movement and Mount St. Helens and flowed south about 11 subsequent buckling of the lava pond crust km in a stream valley incised in pyroclastic prior to, or during, drainage. Some of the flow deposits of late Quaternary age. Two

Figure 29. One of two structures associated with Ole's Cave. The raised-rim crater, about 41 m in Figure 30. Tumulus near the terminus of the Cave diameter, apparently resulted from collapse of a tumulus Basalt. Several domes similar to this occur on the flow; formed in association with the lava tube. A similar some appear to be associated with lava tubes, particularly structure is a few dozen meters down-flow, in line with Lake Cave. The structure is about 50 m in diameter the lava tube. (powerline in foreground).

charcoal samples (from different localities during a temporary damming of the lava flow. within the flow) yield radiocarbon age dates of Although many pressure ridges occur on the 1,860 ± 250 and 1,925 ± 95 years B.P. The flow surface, there is no apparent relation to flow contains sets of lava tubes extending nearly the lava tubes. the entire length of the flow. Most of the tubes apparently formed between shear planes that ACKNOWLEDGMENTS developed within laminar flow of the molten We wish to acknowledge the able field as- lava, similar to the mode of formation described sistance of M. Lovas and K. Oas. We are in- by Oilier and Brown (1965). Where the flow debted to the many spelunkers of the Western gradient increased, however, the active flow Speleological Survey and the National Speleo- probably became turbulent, destroying the logical Society who acted as guides and laminar flow and shear planes, and forming the furnished data and photographs; W. R. Halli- tube roof by accretion of spattered lava on day, C. Larson, C. Baker, and D. Tubbs were lateral levees (for example, Little Red River particularly helpful. The manuscript benefited Cave, upper section) as observed in active lava significantly from critical comments by D. R. flows. Lava tubes and channels of the Cave Crandell, C. A. Hopson, K. G. Snetsinger, Basalt have been modified by subsequent A. W. Hatheway, K. Howard, S. C. Porter, surges of lava. Lack of erosional surfaces within J. F. Kerridge, and W. R. Halliday. Part of the the Cave Basalt indicates that the subsequent data was gathered during a U.S. Geological surges or eruptive phases were all part of the Survey investigation of the potential volcanic same eruptive event. Tube modifications in- hazards of the area by Hyde. Greeley made the clude partial tube filling (Fig. 20), addition of study during tenure as a National Academy of upper tubes to form multiple levels and verti- Sciences-National Research Council Associate. cally elongate cross sections (Fig. 21), and tube Radiocarbon age determinations (except for enlargement by erosion from subsequent surges GX1673) were made in the laboratories of the of lava. U.S. Geological Survey under supervision of Lava flows and tubes are controlled to some Meyer Rubin. degree by preflow topography and are usually This study was supported in part by the situated in topographic depressions. Lava tubes Office of Planetology Programs, National represent the zone of highest flow velocity Aeronautics and Space Administration. within the overall flow body, analogous to a hydrologic thalweg; as such, the axis of the REFERENCES CITED developing tube may migrate in a sinuous Carithers, W., 1946, Pumice and pumicite occur- pattern within active flows. Some sections of rences of Washington: Washington Div. the Cave Basalt tubes, however, appear to Mines and Geology Rept. Inv. 15, 78 p. occupy the bed of a former stream, indicated Crandell, D. R., 1969, Surficial geology of Mount by exposures of country rock where sides of the Rainier National Park, Washington: U.S. tube have collapsed. These exposures also reveal Geol. Survey Bull. 1288, 41 p. that lava flows are capable of eroding preflow Diller, J. C., 1899, Latest volcanic eruptions of the surfaces, shown by undercut sections of country Pacific Coast: Science, new ser., v. 9, p. 239- rock (Fig. 19) and country rock inclusions in 640. the basalt. Elliott, C. P., 1897, Mount St. Helens: Natl. Geog. Mag., v. 8, p. 226-230. Several surface features of basalt flows form Erdmann, C. E., and Warren, W., 1938, Report on in association with lava tubes. A topographic the geology of three dam sites on Toutle ridge often occurs along the lava-tube axis, ap- River, Cowlitz and Skamania Counties, Wash- parently as a result of inflation of the tube and ington: U.S. Geol. Survey Open-File Rept., accretion of lava laterally along the tubes. In 119 p. some places, pressure within the tube was suf- Folsom, M. M., 1970, Volcanic eruptions: The ficient to deform the ceiling into cupolas, or to pioneers' attitude on the Pacific Coast from deform the roof into small surface domes, or 1800 to 1875: The Ore Bin, v. 32, p. 61-71. Forsyth, C. E., 1910, Mount St. Helens: The tumuli. Occasionally, the roof ruptured, Mountaineer, v. 3, p. 56-62. spilling small lava flows onto the surface. Many Gibbs, G., 1854, Report on the geology of the tumuli collapsed to form raised-rim craters. central part of Washington Territory: U.S. Collapse craters near the terminus of Ole's 33d Cong., 1st sess., House Doc. 129, v. 18, pt. Cave may have developed by collapse of hy- 1, p. 494-512. drostatically inflated tumuli over lava tubes Greeley, R., 1969, Geology and morphology of a

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021 2418 GREELEY AND HYDE

small lava tube [abs.]: Am. Geophys. Union Kuiper, G. P., Strom, R. G., and LePoole, R. S., Trans., v. 50, p. 678. 1966, Interpretation of the Ranger records, in 1971a, Lunar Hadley Rille: Considerations of Ranger VIII and IX, Pt. 2: California Inst. its origin: Science, v. 172, p. 722-725. Technology Jet Propulsion Lab. Tech. Rept. 197lb, Observations of actively forming lava 32-800, p. 35-248. tubes and associated structures, Hawaii: Kuno, H., 1968, Differentiation of basalt magma, in Modern Geology, v. 2, p. 207-223. Hess, H. H., and Poldervaart, A., eds., Basalts: 1971c, Geology and morphology of selected The Poldervaart treatise on rocks of basaltic lava tubes in the vicinity of Bend, Oregon: composition: New York, Interscience Pub- Oregon Dept. Geology and Mineral In- lishers, p. 623-688. dustries Bull. 71, 47 p. Lawrence, D. B., 1938, Trees on the march: 1971d, Lava tubes and channels in the lunar Mazama, v. 20, no. 12, p. 49-54. Marius Hills: The Moon, v. 3, p. 289-314. 1939, Continuing research on the flora of Mt. Greeley, R., and Bear, R., 1971, Hambone, Cali- St. Helens: Mazama, v. 21, no. 12, p. 49-59. fornia and its magnificent lava tubes-pre- 1941, The "Floating Island" lava flow of Mt. liminary report: Geol. Soc. America, Abs. with St. Helens: Mazama, v. 23, no. 12, p. 56-60. Programs, v. 3, no. 2, p. 128. 1954, Diagrammatic history of the northeast Greeley, R., and Hyde, J. H., 1970, Lava tubes of slope of Mt. St. Helens, Washington: Mazama, Mount St. Helens, Washington: Geol. Soc. v. 36, no. 13, p. 41-44. America, Abs. with Programs, v. 2, no. 2, p. 96. Mazama, 1903, v. 2, no. 3, p. 134-135. Halliday, W. R., 1963a, Caves of Washington: Mulhneaux, D. R., 1964, Extensive recent pumice Washington Div. Mines and Geology, Inf. lapilh and ash layers from Mount St. Helens Circ. no. 40, p. 71-104. volcano southern Washington [abs.]: Geol. 1963b, Features and significance of the Mount Soc. America Spec. Paper 76, p. 285. St. Helens Cave area: Natl. Parks Mag., v. 37, Mullineaux, D. R., and Crandell, D. R., 1962, no. 195, p. 11-14. Recent lahars from Mount St. Helens, Wash- Hatheway, A. W., 1971, Lava tubes and collapse ington: Geol. Soc. America Bull., v. 73, p. depressions [Ph.D. dissert.]: Univ. Arizona, 855-870. Tucson. Nockolds, S. R., 1954, Average chemical com- Hatheway, A. W., and Herring, A. K., 1970, position of some igneous rocks: Geol. Soc. Bandera lava tubes of New Mexico, and lunar America Bull., v. 65, p. 1007-1052. implications: Arizona Univ. Lunar and Oberbeck, V. R., Quaide, W. L., and Greeley, R., Planetary Lab. Commun., v. 8, pt. 4, p. 298- 1969, On the origin of lunar sinuous rilles: 327. Modern Geology, v. 1, p. 75-80. Holmes, K. L., 1955, Mount St. Helens' recent Oilier, C. D., and Brown, M. C., 1965, Lava caves eruptions: Oregon Hist. Quart., v. 56, p. 197- of Victoria: Bull. Volcanol., v. 28, p. 215-229. 210. Parsons, W. H., 1969, Criteria for the recognition Hopson, C. A., 1971, Eruptive sequence at Mount of volcanic breccias: Review, in Larsen, L., St. Helens, Washington: Geol. Soc. America, Prinz, M., and Manson, V., eds., Igneous and Abs. with Programs, v. 3, p. 138. metamorphic geology: Geol. Soc. America Hyde, J. H., 1970, Geologic setting of Merrill Lake Mem. 115, p. 270. and evaluation of volcanic hazards in the Pryde, P. R., 1968, Mount Saint Helens: A possible Kalama River vaU.ey near Mount St. Helens, national monument: Natl. Parks Mag., v. 42, Washington: U.S. Geol. Survey Open-File no. 248, p. 7-10. Rept., 17 p. Verhoogen, J., 1937, Mount St. Helens: A recent Hyde, J. H., and Greeley, R., 1971, Phreatic ex- cascade volcano: California Univ., Dept. Geol. plosion in a lava tube: Am. Geophys. Union Sci. Bull., v. 24, no. 9, p. 263-302. Trans., v. 52, p. 433. Waters, A. C., 1961, Stratigraphic and lithologic Jillson, W. R., 1917a, The volcanic activity of variations in the Columbia River Basalt: Am. Mount St. Helens and Mount Hood in histori- Jour. Sci., v. 259, p. 583-611. cal time: Geog. Rev., v. 3, p. 481-485. Williams, I. A., 1922, Tree casts in recent lava: 1917b, New evidence of a recent volcanic Natural History, v. 22, no. 6, p. 543-548. eruption on Mt. St. Helens, Washington: Am. Jour. Sci., v. 44, no. 259, p. 59-62. 1921, Physiographic effects of the volcanism MANUSCRIPT RECEIVED BY THE SOCIETY AUGUST of Mt. St. Helens: Geog. Rev., v. 11, p. 398- 26, 1971 405. REVISED MANUSCRIPT RECEIVED MARCH 9, 1972

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/83/8/2397/3443198/i0016-7606-83-8-2397.pdf by guest on 02 October 2021