THE RISE OF MOLTEN ROCK TO THE EARTH’S SURFACE

VOLCANIC MOUNTAINS OF EXUDATION

Prevalent misconceptions about volcanoes.—The more or less common impression that a volcano is a “burning mountain” or a “smoking mountain” has been much fostered by the school texts in physical geography in use during an earlier period. The best introduction to a discussion of volcanoes is, therefore, a disillusionment from this notion. Far from being burning or smoking, there is normally no combustion whatever in connection with a volcanic eruption. The unsophisticated tourist who, looking out from Naples, sees the steam cap which overhangs the Vesuvian crater tinged with brown, easily receives the impression that the material of the cloud is smoke. Even more at night, when a bright glow is reflected to his eye and soon fades away, only to again glow brightly after a few moments have passed, is it difficult to remove the impression that one is watching an intermittent combustion within the crater. The cloud which floats away from the crest of the mountain is in reality composed of steam with which is admixed a larger or smaller proportion of fine rock powder which gives to the cloud its brownish tone. The glow observed at night is only a reflection from molten lava within the crater, and the variation of its brightness is explained by the alternating rise and fall of the lava surface by a process presently to be explained.

Not only is there no combustion in connection with volcanic eruptions, but so far as the volcano is a mountain it is a product of its own action. The grandest of volcanic eruptions have produced no mountains whatever, but only vast plains or plateaus of consolidated molten rock, and every volcanic mountain at some time in its history has risen out of a relatively level surface.

When the traditional notions about volcanoes grew up, it was supposed that the solid earth was merely a “crust” enveloping still molten material. As has already been pointed out in an earlier chapter, this view is no longer tenable, for we now know that the condition of matter within the earth’s interior, while perhaps not directly comparable to any that is known, yet has properties most resembling known matter in a solid state; it is much more rigid than the best tool steel. While there must be reservoirs of molten rock beneath active volcanoes, it is none the less clear that they are small, local, and temporary. This is shown by the comparative study of volcanic outlets within any circumscribed district.

It is perhaps not easy to frame a definition of a volcano, but its essential part, instead of being a mountain, is rather a vent or channel which opens up connection between a subsurface reservoir of molten rock and the surface of the earth. An eruption occurs whenever there is a rise of this material, together with more or less steam and admixed gases, to the surface. Such molten rock arriving at the surface is designated lava. The changes in pressure upon this material during its elevation induce secondary phenomena as the surface is approached, and these manifestations are often most awe inspiring. While often locally destructive, the geological importance of such phenomena is by reason of their terrifying aspect likely to be greatly exaggerated.

Early views concerning volcanic mountains.—As already pointed out, a volcano at its birth is not a mountain at all, but only, so to speak, a shaft or channel of communication between the surface and a subterranean reservoir of molten rock. By bringing this melted rock to the surface there is built up a local elevation which may be designated a mountain, except where the volume of the material is so large and is spread to such distances as to produce a plain (see fissure eruptions below).

In the early history of geology it was the view of the great German geologist von Buch and his friend and colleague von Humboldt, that a volcanic mountain was produced in much the same manner as is a blister upon the body. The fluids which push up the cuticle in the blister were here replaced by fluid rock which elevated the sedimentary rock layers at the surface into a dome or mound which was open at the top—the so-called crater. This “elevation-crater” theory of volcanoes long held the stage in geological science, although it ignored the very patent fact that the layers on the flanks of volcanic cones are not of sedimentary rock at all, but, on the contrary, of the volcanic materials which are brought up to the surface during the eruption. The observational phase of science was, however, dawning, and the English geologists Scrope and Lyell were able to show by study of volcanic mountains that the mound about the volcanic vent was due to the accumulation of once molten rock which had been either exuded or ejected. Making use of data derived from New Zealand, Scrope showed that, instead of being elevated during the formation of a volcanic mountain, the sedimentary strata of the vicinity may be depressed near the volcanic vent (Fig. 87).

The birth of volcanoes.—To confirm the impression that the formation of the volcanic mountain is in reality a secondary phenomenon connected with eruptions, we may cite the observed birth of a number of volcanoes. On the 20th of September, 1538, a new volcano, since known as Monte Nuovo (new mountain), rose on the border of the ancient Lake Lucrinus to the westward of Naples. This small mountain attained a height of 440 feet, and is still to be seen on the shore of the bay of Naples. From Mexico have been recorded the births of several new volcanoes: Jorullo in 1759, Pochutla in 1870, and in 1881 a new volcano in the Ajusco Mountains about midway between the Gulf of Mexico and the Pacific Ocean. The latest of new volcanoes is that raised in Japan on November 9, 1910, in connection with the eruption of Usu-san. This “New Mountain” reached an elevation of 690 feet.

As described by von Humboldt, Jorullo rose in the night of the 28th of September, 1759, from a fissure which opened in a broad plain at a point 35 miles distant from any then existing volcano. The most remarkable of new volcanoes rose in 1871 on the island of Camiguin northward from Mindanao in the Philippine archipelago. This mountain was visited by the Challenger expedition in 1875, and was first ascended and studied thirty years later by a party under the leadership of Professor Dean C. Worcester, the Secretary of the Interior of the Philippine Islands, to whom the writer is indebted for this description and the accompanying illustration of this largest and most interesting of new-born volcanoes. As in the case of Jorullo, the eruption began with the formation of a fissure in a level plain, some 400 yards distant from the town of Catarman (Fig. 88). The eruption continued for four years, at the end of which time the height of the summit was estimated by the Challenger expedition to be 1900 feet. At the time of the first ascent in 1905, the height was determined by aneroid as 1750 feet, with sharp rock pinnacles projecting some 50 or 75 feet higher.

Active and extinct volcanoes.—The terms “active” and “extinct” have come into more or less common use to describe respectively those volcanoes which show signs of eruptive activity, and those which are not at the time active. The term “dormant” is applied to volcanoes recently active and supposed to be in a doubtfully extinct condition. From a well-known volcano in the vicinity of Naples, volcanoes which no longer erupt lava or cinder, but show gaseous emanations (fumeroles) are said to be in the solfatara condition, or to show solfataric activity.

Experience shows that the term “extinct”, while useful, must always be interpreted to mean apparently extinct. This may be illustrated by the history of Mount Vesuvius, which before the Christian era was forested in the crater and showed no signs of activity; and in fact it is known that for several centuries no eruption of the volcano had taken place. Following a premonitory earthquake felt in the year 63, the mountain burst out in grand explosive eruption in 79 A.D. This eruption profoundly altered the aspect of the mountain and buried the cities of Pompeii, Stabeii, and Herculaneum from sight. Once more, this time during the middle ages, for nearly five centuries (1139 to 1631) there was complete inactivity, if we except a light ash eruption in the year 1500. During this period of rest the crater was again forested, but the repose was suddenly terminated by one of the grandest eruptions in the mountain’s history.

The earth’s volcano belts.—The distribution of volcanoes is not uniform, but, on the contrary, volcanic vents appear in definite zones or belts, either upon the margins of the continents or included within the oceanic areas (Fig. 89). The most important of these belts girdles the Pacific Ocean, and is represented either by chains or by more widely spaced volcanic mountains throughout the Cordilleran Mountain system of South and Central America and Mexico, by the volcanoes of the Coast and Cascade ranges of North America, the festooned volcanic chain of the Aleutian Islands, and the similar island arcs off the eastern coast of the Eurasian continent. The belt is further continued through the islands of Malaysia to New Zealand, and on the Pacific’s southern margin are found the volcanoes of Victoria Land, King Edward Land, and West Antarctica.

This volcano girdle is by no means a perfect one, for in addition to the principal festoons of the western border there are many secondary ones, and still other arcs are found well toward the center of the oceanic area. Another broad belt of volcanoes borders the Mediterranean Sea, and is extended westward into the Atlantic Ocean. Narrower belts are found in both the northern and southern portions of the Atlantic Ocean, on the margins of the Caribbean Sea, etc. The fact of greatest significance in the distribution seems to be that bands of active volcanoes are to be found wherever mountain ranges are paralleled by deeps on the neighboring ocean floor (Fig. 90). As has been already pointed out in the chapter upon earthquakes, it is just such places as these which are the seat of earthquakes; these are zones of the earth’s crust which are undergoing the most rapid changes of level at the present time. Thus the rise of the land in mountains is proceeding simultaneously with the sinking of the sea floor to form the neighboring deeps.

Arrangement of volcanic vents along fissures and especially at their intersections.—Within those districts in which volcanoes are widely separated from their neighbors, the law of their arrangement is difficult to decipher, but the view that volcanic vents are aligned over fissures is now supported by so much evidence that illustrations may be supplied from many regions. An exceptionally perfect line of small cones is found along the SkaptÁr cleft in Iceland, upon which stands the large volcano of Laki. This fissure reopened in 1783, and great volumes of lava were exuded. Over the cleft there was left a long line of volcanic cones (Fig. 91). There are in Iceland two dominating series of parallel fissures of the same character which take their directions respectively northeast-southwest and north-south. Many such fissures are traceable at the surface as deep and nearly straight clefts or gjÁs, usually a few yards in width, but extending for many miles. The EldgjÁ has a length of more than 18 English miles and a depth varying from 400 to 600 feet. On some of these fissures no lava has risen to the surface, whereas others have at numerous points exuded molten rock. Sometimes one end only of a fissure, the more widely gaping portion, has supplied the conduits for the molten lava. This is well illustrated by the cratered monticules raised by the common ant over the cracks which separate the blocks of cement sidewalk, the hillocks being located where the most favorable channel was found for the elevation of the materials.

Those places upon fissures which become lava conduits appear to be the ones where the cleft gapes widest so as to furnish the widest channel. Wherever a differential lateral movement of the walls has occurred, openings will be found in the neighborhood of each minor variation from a straight line (Fig. 92 a). Wherever there are two or more series of fissures, and this would appear to be the normal condition, places favorable for lava conduits occur at fissure intersections. Within such veritable volcano gardens as are to be found in Malaysia, the law of volcano distribution became apparent so soon as accurate maps had been prepared. Thus the outline map of a portion of the island of Java (Fig. 93) shows us that while the volcanoes of the island present at first sight a more or less irregular band or zone, there are a number of fissures intersecting in a network, and that the volcanoes are aligned upon the fissures with the larger cones located at the intersections. So also in Iceland, the great eruption of Askja in 1875 occurred at the intersection of two lines of fissure.

Outside these closely packed volcanic regions, similar though less marked networks are indicated; as, for example, in and near the Gulf of Guinea. If now, instead of reducing the scale of our volcano maps, we increase it, the same law of distribution is no less clearly brought out. The monticules or small volcanic cones which form upon the flanks of larger volcanic mountains are likewise built up over fissures which on numerous occasions have been observed to open and the cones to form upon them.

Still further reducing now the area of our studies and considering for the moment the “frozen” surface of the boiling lava within the caldron of Kilauea, this when observed at night reveals in great perfection the sudden formation of fissures in the crust with the appearance of miniature volcanoes rising successively at more or less regular intervals along them.

It not infrequently happens that after a volcanic vent has become established above some conduit in a fissure, the conduit migrates along the fissure, thus establishing a new cone with more or less complete destruction of the old one (Fig. 94).

The so-called fissure eruptions.—The grandest of all volcanic eruptions have been those in which the entire length and breadth of the fissures have been the passageway for the upwelling lava. Such grander eruptions have been for the most part prehistoric, and in later geologic history have occurred chiefly in India, in Abyssinia, in northwestern Europe, and in the northwestern United States. In western India the singularly horizontal plateaus of basaltic lava, the Dekkan traps, cover some 200,000 square miles and are more than a mile in depth. The underlying basement where it appears about the margins of the basalt is in many places intersected by dikes or fissure fillings of the same material. No cones or definite vents have been found.

The larger portion of the northwestern British Isles would appear to have been at one time similarly blanketed by nearly horizontal beds of basaltic lava, which beds extended northwestward across the sea through the Orkney and Faroe islands to Iceland. Remnants of this vast plateau are to-day found in all the island groups as well as in large areas of northeastern Ireland, and fissure fillings of the same material occur throughout large areas of the British Isles. In many cases these dikes represent once molten rock which may never have communicated with the surface at the time of the lava outpouring, yet they well illustrate what we might expect to find if the basalt sheets of Iceland or Ireland were to be removed.

The floods of basaltic lava which in the northwestern United States have yielded the barren plateau of the Cascade Mountains (Fig. 95) would appear to offer another example of fissure eruption, though cones appear upon the surface and perhaps indicate the position of lava outlets during the later phases of the eruptive period. The barrenness and desolation of these lava plains is suggested by Fig. 96.

Though the greater effusions of lava have occurred in prehistoric times, and the manner of extrusion has necessarily been largely inferred from the immense volume of the exuded materials and the existence of basaltic dikes in neighboring regions, yet in Iceland we are able to observe the connection between the dikes and the lava outflows. Professor Thoroddsen has stated that in the great basaltic plateau of Iceland, lava has welled out quietly from the whole length of fissures and often on both sides without giving rise to the formation of cones. At three wider portions of the great Eld cleft, lava welled out quietly without the formation of cones, though here in the southern prolongation of the fissure, where it was narrower, a row of low slag cones appeared. Where the lava outwellings occurred, an area of 270 square miles was flooded.

The composition and the properties of lava.—In our study of igneous rocks (Chapter IV) it was learned that they are composed for the most part of silicate minerals, and that in their chemical composition they represent various proportions of silica, alumina, iron, magnesia, lime, potash, and soda. The more abundant of these constituents is silica, which varies from 35 to 70 per cent of the whole. Whenever the content of silica is relatively low,—basic or basaltic lava,—the cooled rock is dark in color and relatively heavy. It melts at a relatively low temperature, and is in consequence relatively fluid at the temperatures which lavas usually have on reaching the earth’s surface. Furthermore, from being more fluid, the water which is nearly always present in large quantity within the lava more readily makes its escape upon reaching the surface. Eruptions of such lava are for this reason without the violent aspects which belong to extrusions of more siliceous (more “acidic”) lavas. For the same reason, also, basaltic lava flows more freely and can spread much farther before it has cooled sufficiently to consolidate. This is equivalent to saying that its surface will assume a flatter angle of slope, which in the case of basaltic lava seldom exceeds ten degrees and may be less than one degree (Fig. 97).

Siliceous lavas, on the other hand, are, when consolidated, relatively light both in color and weight and melt at relatively high temperatures. They are, therefore, usually but partly fused and of a viscous consistency when they arrive at the earth’s surface. Because of this viscosity they offer much resistance to the liberation of the contained water, which therefore is released only to the accompaniment of more or less violent explosions. The lava is blown into the air and usually falls as consolidated fragments of various degrees of coarseness.

It must not, however, be assumed that the temperature of lava is always the same when it arrives at the surface, and hence it may happen that a siliceous lava is exuded at so high a temperature that it behaves like a normal basaltic lava. On the other hand, basaltic lavas may be extruded at unusually low temperatures, in which case their behavior may resemble that of the normal siliceous lavas. If, however, as is generally the case, the energy of explosion of a basaltic lava is relatively small, any ejected portions of the liquid lava travel to a moderate height only in the air, so that on falling they are still sufficiently pasty to adhere to rock surfaces and thus build up the remarkably steep cones and spines known as “spatter cones” or “driblet cones” (Fig. 98). When, on the other hand, the energy of explosion is great, as is normally the case with siliceous lavas, the portions of ejected lava have been fully consolidated before their fall to the surface, so that they build up the same type of accumulation as would sand falling in the same manner. The structures which they form are known as tuff, cinder, or ash cones (Fig. 99).

Whenever the contained water passes off from siliceous lavas without violent explosions, the lava may flow from the vent, but in contrast to basaltic lavas it travels a short distance only before consolidating. The resulting mountain is in consequence proportionately high and steep (Fig. 97). Eruptions characterized by violent explosions accompanied by a fall of cinder are described as explosive eruptions. Those which are relatively quiet, and in which the chief product is in the form of streams of flowing lava, are spoken of as convulsive eruptions.

The three main types of volcanic mountain.—If the eruptions at a volcanic vent are exclusively of the explosive type, the material of the mountain which results is throughout tuff or cinder, and the volcano is described as a cinder cone. If, on the other hand, the vent at every eruption exudes lava, a mountain of solid rock results which is a lava dome. It is, however, the exception for a volcano which has a long history to manifest but a single kind of eruption. At one time exuding lava comparatively quietly, at another the violence with which the steam is liberated yields only cinder, and the mountain is a composite of the two materials and is known as a composite volcanic cone.

The lava dome.—When successive lava flows come from a crater, the structure which results has the form of a more or less perfect dome. If the lava be of the basaltic or fluid type, the slopes are flat, seldom making an angle of as much as ten degrees with the horizon and flatter toward the summit (Fig. 101, p. 106). If of siliceous or viscous lava, on the other hand, the slopes are correspondingly steep and in some cases precipitous. To this latter class belong some of the Kuppen of Germany, the puys of central France, and the mamelons of the Island of Bourbon.

The basaltic lava domes of Hawaii.—At the “crossroads of the Pacific” rises a double line of lava volcanoes which reach from 20,000 to 30,000 feet above the floor of the ocean, some of them among the grandest volcanic mountains that are known. More than half the height and a much larger proportion of the bulk of the largest of these are hidden beneath the ocean’s surface. The two great active vents are Mokuaweoweo (on Mauna Loa) and Kilauea, distinct volcanoes notwithstanding the fact that their lava extravasations have been merged in a single mass. The rim of the crater of Mauna Loa is at an elevation of 13,675 feet above the sea, whereas that of Kilauea is less than 4000 feet and appears to rest upon the flank of the larger mountain (Figs. 100 and 101). Although one crater is but 20 miles distant from the other and nearly 10,000 feet lower, their eruptions have apparently been unsympathetic. Nowhere have still active lava mountains been subjected to such frequent observations extending throughout a long period, and the dynamics of their eruptions are fairly well understood. To put this before the reader, it will be best to consider both mountains, for though they have much in common, the observations from one are strangely complementary to those of the other. The lower crater being easily accessible, Kilauea has been often visited, and there exists a long series of more or less consecutive observations upon it, which have been assembled and studied by Dana and Hitchcock. The place of outflow of the Kilauea lavas has not generally been visible, whereas Mokuaweoweo has slopes rising nearly 14,000 feet above the sea and displays the records of outflow of many eruptions, some of which were accompanied by the grandest of volcanic phenomena.

Lava movements within the caldron of Kilauea.—The craters of these mountains are the largest of active ones, each being in excess of seven miles in circumference. In shape they are irregularly elliptical and consist of a series of steps or terraces descending to a pit at the bottom, in which are open lakes of boiling lava. Enough is known of the history of Kilauea to state that the steep cliffs bounding the terraces are fault walls produced by inbreak of a frozen lava surface. The cliff below the so-called “black ledge” was produced by the falling in of the frozen lava surface at the time of the outflow of 1840, the lava issuing upon the eastern flank of the mountain and pouring into the sea near Nanawale. Since that date the floor of the pit below the level of this ledge has been essentially a movable platform of frozen lava of unknown and doubtless variable thickness which has risen and descended like the floor of an elevator car between its guiding ways (Fig. 102). The floor has, however, never been complete, for one or more open lakes are always to be seen, that of Halemaumau located near the southwestern margin having been much the most persistent. Within the open lakes the boiling lava is apparently white hot at the depth of but a few inches below the surface, and in the overturnings of the mass these hotter portions are brought to the surface and appear as white streaks marking the redder surface portions. From time to time the surface freezes over, then cracks open and erupt at favored points along the fissures, sending up jets and fountains of lava, the material of which falls in pasty fragments that build up driblet cones. Small fluid clots are shot out, carrying a threadlike line of lava glass behind them, the well-known “PelÉ’s hair.” Sometimes the open lakes build up congealed walls, rising above the general level of the pit, and from their rim the lava spills over in cascades to spread out upon the frozen floor, thus increasing its thickness from above (Fig. 103). At other times a great dome of lava has been pushed up from the pit of Halemaumau under a frozen shell, the molten lava shining red through cracks in its surface and exuding so as to heal each widely opened fissure as it forms.

At intervals of from a few years to nine or ten years the crater has been periodically drained, at which times the moving platform of frozen lava has sunk more or less rapidly to levels far below the black ledge and from 900 to 1700 feet below the crater rim. Following this descent a slow progressive rise is inaugurated, which has sometimes gone on at a rate of more than a hundred feet per year, though it is usually much slower than this. When the platform has reached a height varying from 700 to 350 feet below the crater rim, another sudden settlement occurs which again carries the pit floor downward a distance of from 300 to 700 feet.

The draining of the lava caldrons.—The changes which go on within the crater of Mokuaweoweo, though less studied than those of Kilauea, appear to be in some respects different. Here every eruption seems to be preceded by a more or less rapid influx of melted lava to the pit of the crater, this phenomenon being observed from a distance as a brilliant light above the crater—the reflection of the glow from overhanging vapor clouds. The uprising of the lava has often been accompanied by the formation of high lava fountains upon the surface, and the molten lava sometimes appears in fissures near the crater rim at levels well above the lava surface within the pit.

Although in many cases the lava which has thus flooded the crater has suddenly drained away without again becoming visible, it is probable that in such cases an outlet has been found to some submarine exit, since under-ocean discharge effects have been observed in connection with eruptions of each of the volcanoes.

Inasmuch as no earthquakes are felt in connection with such outflows as have been described, it is probable that the hot lava fuses a passageway for itself into some open channel underneath the flanks of the mountain. Such a course is well illustrated by the outflow of Kilauea in 1840, when, it will be remembered, occurred the great down-plunge of the crater that yielded the pit below the black ledge. At this time the lava first made its appearance upon the flanks of the mountain at the bottom of a small pit or inbreak crater which opened five miles southeast of the main crater of Kilauea (Fig. 104). Within this new crater the lava rose, and small ejections soon followed from fissures formed in its neighborhood. Some time after, the lava sank in the first new crater, only to reappear successively at other small openings (Fig. 104, B, C, m, n) and finally to issue in volume at a point eleven miles from the shore and flow thereafter upon the surface of the mountain until it had reached the sea. Only the slightest earth tremors were felt, and as no rumblings were heard, it is evident that the lava fused its way along a buried channel largely open at the time (see below, p. 112).

In a majority of the eruptions of Mokuaweoweo, when the outflowing lavas have become visible, the molten rock has apparently fused its way out to the surface of the mountain at points from 1000 to 3000 feet below the bottom of the crater, and this discharge has corresponded in time to the lowering of the lava surface within the crater. There are, however, three instances upon record in which the lava issued from definite rents which were formed upon the mountain flanks at comparatively low levels. In contrast to the formation of fused outlets, these ruptures of a portion of the mountain’s flank were always accompanied by vigorous local earthquakes of short duration. In one instance (the eruption of 1851) such a rent appeared under the same conditions but at an elevation of 12,500 feet, or near the level of the lava in the crater.

The outflow of the lava floods.—In order to properly comprehend these and many otherwise puzzling phenomena connected with volcanoes, it is necessary to keep ever in mind the quite remarkable heat-insulating property of congealed lava. So soon as a thin crust has formed upon the surface of molten rock, the heat of the underlying fluid mass is given off with extreme slowness, so that lava streams no longer connected with their internal lava reservoirs may remain molten for decades.

We have seen that for Mokuaweoweo and Kilauea, lava either quietly melts its way to the surface at the time of outflow, or else produces a rent for its egress to the accompaniment of vigorous local earthquakes. In either case if the lava issues at a point far below the crater, gigantic lava fountains arise at the point of outflow, the fluid rock shooting up to heights which range from 250 to 600 or more feet above the surface. A certain proportion of this fluid lava is sufficiently cooled to consolidate while traveling in the air, and falling, it builds up a cinder cone which is left as a location monument for the place of discharge. From this outlet the molten lava begins its journey down the slope of the mountain, and quickly freezes over to produce a tunnel, beneath the roof of which the fluid lava flows with comparatively slow further loss of heat. Save for occasional steam jets issuing from its surface, it may give little indication of its presence until it has reached the sea (Fig. 105).

If sufficient in volume and the shore be not too distant, the stream of lava arrives at the sea, where, discharging from the mouth of its tunnel, it throws up vast volumes of steam and induces ebullition of the water over a wide area (Fig. 106). Professor Dana, who visited Hawaii a few months only after the great outflow of 1840, states that the lava, upon reaching the ocean, was shivered like melted glass and thrown up in millions of particles which darkened the sky and fell like hail over the surrounding country. The light was so bright that at a distance of forty miles fine print could be read at midnight.

Protected from any extensive consolidation by its congealed cover, the lava within a stream may all drain away, leaving behind an empty lava tunnel, which in the case of the Hawaiian volcanoes sometimes has its roof hung with beautiful lava stalactites and its floor studded with thin lava spines. Later lava outflows over the same or neighboring courses bury such tunnels beneath others of similar nature, giving to the mountain flanks an elongated cellular structure illustrated schematically in Fig. 107. These buried channels may in the future be again utilized for outflows similar in character to that of Kilauea in 1840.

While the formation of lava stalactites of such perfection and beauty is peculiar to the Hawaiian lava tunnels, the formation of the tunnel in connection with lava outflow is the rule wherever a dissipation at the end has permitted of drainage. A few hours only after the flow has begun, the frozen surface has usually a thickness of a few inches, and this cover may be walked over with the lava still molten below. At first in part supported by the molten lava, the tunnel roof sometimes caves in so soon as drainage has occurred.

Wherever basaltic lava has spread out in valleys on the surface of more easily eroded material, either cinder or sedimentary formations, the softer intervening ridges are first carried away by the eroding agencies, leaving the lava as cappings upon residual elevations. Thus are derived a type of table mountain or mesa of the sort well illustrated upon the western slopes of the Sierra Nevadas in California (Fig. 108).

The surface which flowing lava assumes, while subject to considerable variation, may yet be classified into two rather distinct types. On the one hand there is the billowy surface in which ellipsoidal or kidney-shaped masses, each with dimensions of from one to several feet, lie merged in one another, not unlike an irregular collection of sofa pillows. This type of lava has become known as the Pahoehoe, from the Hawaiian occurrence (Fig. 109). A variation from this type is the “corded” or “ropy” lava, the surface of which much resembles rope as it is coiled along the deck of a vessel, the coils being here the lines of scum or scoriÆ arranged in this manner by the currents at the surface of the stream (Fig. 123, p. 124). A quite different type is the block lava (Aa type) which usually has a ragged scoriaceous surface and consists of more or less separate fragments of cooled lava (Fig. 131, p. 130).

Wherever lava flows into the sea in quantity, it extends the margin of the shore, often by considerable areas. The outflow of Kilauea in 1840 extended the shore of Hawaii outward for the distance of a quarter of a mile, and a more recent illustration of such extension of land masses is furnished by Fig. 110.

---