The Main Pumice Fall

     Lithology of the Main Pumice Fall

Next to be considered are the variations in the coarseness of the ejecta, the proportion of crystals to glass, and the amount of lithic detritus, that is, fragments of old lava, present among the deposits.

Sorting of the pumice by size. Moore⁵ rendered a valuable service in preparing sixty-one mechanical analyses of samples of the pumice fall, a few of which are reproduced here as figure 18. Many more separations were made during the present study, and some are included in figures 17, 19, 20, 21, and 22.

In a general way, as might be expected, the pumice becomes finer away from Crater Lake, though the more the deposits are examined the more clearly it ears that this is indeed no more than a generalization. Commonly there is much more variation in a vertical thickness of a few feet than there is in many miles on the surface. Since, moreover, layers of a given coarseness and sorting can only be correlated over small areas; samples taken at random depths give a misleading picture. It should also be emphasized that the average coarseness does not by any means vary with the thickness of the deposits. Along the western fringe of the pumice fall, even where the thickness is less than 6 inches, lumps as much as 3 inches across are far from rare, though in other places nearer the source, where the thickness exceeds 2 feet, lumps more than an inch in diameter are exceptional.

Let it be repeated, however, that in general the pumice becomes finer with increasing distance from Crater Lake. Thus, as the histograms (figure 19) show, approximately two-thirds of the pumice that fell on Timber Crater, 8 miles from the center of the lake, consists of fragments more than 10 mm. in diameter. On the west slope of Mount Thielsen, 14 miles from the source, the percentage of such fragments is 62; near Miller Lake, on the east side of Thielsen, 21 miles from the source, the percentage drops to 25; near Windigo Pass, 25 miles from the source, it is 15; and on Skookum Butte, 30 miles away, the percentage is 18. Still farther from Crater Lake, the proportion of large lumps diminishes more rapidly. In the samples represented by figure 21, hardly any of the ejecta reach 10 mm. across.

The largest pumice lumps are to be found on Timber Crater, where many exceed 6 inches in maximum dimension, and a few reach a length of 8 inches. Lumps up to 4 1/2 inches across may be collected as far as 30 miles from the source, on Skookum Butte and Walker Rim. Even 40 miles to the northeast there are occasional lumps 3 inches across. Seventy miles distant in the same direction, pieces between 1/2 and I inch in size are not uncommon. Clearly, most of the large lumps were drifted to the north and northeast, and were products of the later eruptions, just preceding the pumice flows. During the earlier and weaker activity, the winds blew eastward, and therefore the pumice in that direction is much finer. Even on the hills near Kirk, lumps more than 2 inches across are far from abundant. On Boundary and Lookout buttes a few pieces measure 3 inches across, but around Wocus Bay, between 20 and 32 miles from Crater Lake, the percentage of lumps more than 1/2 inch in diameter ranges between zero and 13 (see histograms, figure 20).

If all the pumice fragments were equally vesicular, the products of a single eruption would tend to diminish in size away from the vent. The fact is, however, that the porosity of the pumice varies considerably; some pieces are extremely cellular, whereas others are scarcely "frothed" at all. Large, light lumps may therefore be found among much smaller and more compact ejecta. Local variations in the content of large lumps must also have resulted from changes in the intensity and direction of explosions.

Generally, at any given locality the pumice becomes coarser from the bottom upward. This reflects the growing violence of the eruptions. When the explosions began to hurl out lumps between 2 and 3 inches across, the winds from the west began to veer toward the southwest. Hence in the region east and southeast of Klamath Marsh, where the deposits are more than 10 feet thick, the proportion of lumps greater than 3 inches is extremely small; farther north the proportion is high even where the deposits are only half as thick. At no time did large lumps fall south of Crater Lake. Even 2 miles south of the caldera, on the Rim Road, where the deposit is 8 feet thick, few lumps measure more than 2 inches across.

Compared with the deposits of the pumice flows, the material of the pumice fall is much better sorted (see histograms, figure 24). In view of the mode of deposition, this is not surprising, for the material of the pumice fall was subjected to long transport in the air and fell largely according to the dictates of gravity.

The deposits of the pumice flows, on the other hand, ere laid down rapidly in the manner of avalanches and therefore suffered no sorting by wind.

Particularly striking is the paucity of fine dust in the pumice fall and its abundance in the pumice flows. Samples taken from the snout of the pumice flow poured down the Rogue River contain no less 70 per cent by volume of material finer than 0.5 mm. Elsewhere, more than half the flows are made up of particles less than 1 mm. in diameter. On the contrary, the proportion of fine dust is insignificant in the deposits of the pumice fall, particularly close to Crater Lake (see figure 19). Less than I per cent of samples taken from Timber Crater, Mount Thielsen, Miller Lake, and Windigo Pass consist of ejecta smaller than 1 mm. in size. Farther away, the content of fine particles increases rapidly (figure 21).

In the analyses listed by Moore, the contrast between the "fines" in the flows and in the fall is very clear. Considering only the fraction less than 64 mm. in size, the averages are as follows:

In other words, almost two-thirds of the pumice flow material below 64 mm. in diameter measures less than I mm, though in the pumice fall less than a quarter is below that size. Furthermore, by far the bulk of the fine dust in the flows is pulverized pumiceous glass, whereas much of the dust in the fall consists of broken crystals. Many samples collected in the area south and east of Klamath Marsh contain merely a trace of material less than I mm. in size, and this may consist entirely of crystals. Even in the finest pumice fall examined, from the Paisley Cave, 80 miles east of Crater Lake, the percentage of dust less than 0.25 mm. in size is only 3.

The reason for the much greater content of "fines" in the pumice flows is clearly related to the comminution of the material as it was carried along in the form of avalanches, the fragments continually bombarding one another as the glowing masses swept down the sides of the volcano. Among the particles in the pumice fall there was, of course, no such mutual attrition.

Nature of the pumice in the fall. Close to the surface, where the pumice has been weathered, the color is pale b d or brownish, but below it is usually white ok pale gray. Where vegetation is heavy, the zone of weathering may be as much as 2 feet thick, but generally it is only a few inches. Soil is either extremely thin or altogether absent.

Most of the fragments are approximately equidimensional, but many are spindle- or disk-shaped and some are drawn out into long threads. In certain fragments the vesicles are ovoid; in others they are tubular, so that the pumice has a shredded, fibrous appearance. In the larger lumps the vesicles occupy much space, but they become progressively smaller as the size of the lumps diminishes, and some of the finest dust is made up of glass shards almost wholly devoid of pores. The difference results, of course, from the bursting of gas bubbles in the larger clots. Accordingly, small shards of glass may weigh as much as cellular fragments many times larger, and for that reason ejecta of quite different sizes settled from the air simultaneously over the same regions.

Unlike the large lumps in the pumice flows, which were rounded by abrasion during transport, the larger fragments in the pumice fall are either angular or subangular. Many of the larger pieces are pink throughout; many more show a narrow pink zone a few millimeters from the surface. Apparently the pink color results from atmospheric oxidation of iron-bearing gases given off by the pumice in a heated condition. The smaller pieces fail to show the color because they were chilled too rapidly and ceased almost at once to give off gas.

Though most, of the larger pumice fragments are extremely vesicular, there are a few, particularly within 20 miles of Crater Lake, which are dark gray, compact, and lithoidal. These represent clots of magma chilled quickly, probably prior to explosion.

Crystal content. Casual inspection gives the false impression that crystals form only a small fraction of the pumice fall. When the material is screened and examined with a hand lens, crystals are seen to be abundant. In size, they range from about 0.2 to 2 mm, by far the majority measuring between 0.5 and I mm. in longest dimension. In most samples, the fraction between those limits may be almost exclusively and is usually at least one-third crystals. In the larger fragments of pumice many crystals are still embedded in the glass, but in pieces smaller than a few millimeters across the separation of glass from crystals is almost complete. This separation is not the result of weathering and disintegration, but an original charmer, caused by liberation of the crystals from the enclosing droplets of viscous glass as they spun through the air.

The ratio of crystals to glass varies between wide limits. Close to Crater Lake, where the deposits are crowded with large lumps of pumice, the volume percentage of discrete crystals commonly falls below 1. At greater distances the percentage may rise above 25, and in one sample, from Wikiup on the Deschutes River, it reaches 40. Surprising was the discovery that among the deposits bordering the Dalles-California highway, the content of crystals generally increases with the distance from Crater Lake (figure 21). In the region southeast of the Klamath Marsh, the volume percentage of crystals ranges from I to 11 (figure 20).

Because of the great lateral and vertical variations, and the difficulty of estimating the amount of crystals still embedded in glass, it is impossible to give an accurate figure for the total content of crystals in the fall. Moore estimated the volume percentage to be about 5, but this figure seems too low. Probably 10 to 15 per cent is nearer the truth.

As to the nature of the crystals, they consist of plagioclase, hypersthene, and smaller amounts of augite and hornblende. The ratios of heavy to light minerals are shown in some of the histograms (figures 19-22). One might expect that toward the source the proportion of heavy minerals would increase at the expense of the light, but no such regular variation was observed. The only generalization which can be made is that in the marginal parts of the pumice sheet the ratio of light to heavy crystals is usually higher than it is nearer the source. In some samples feldspar may be ten times as abundant as the dark minerals, but the usual ratio is between 2 and 3 to I. Being in general both stouter and larger, the feldspar crystals, despite their lower density, tended to fall at the same time as the heavier crystals. Changes in wind velocity, in the strength of the eruptions, and in the crystal content of the magma combine to explain the irregular variations noted above.

If our estimate of the average content of crystals, between 10 and 15 per cent, be correct, it follows that the percentage in the magma prior to explosion was much larger, for when the magma frothed into pumice its volume was greatly increased. In other words, just before the pumice eruptions the upper part of the magma chamber had crystallized to the extent of more than 30 per cent of its volume. Retrograde boiling associated with this advanced state of crystallization may well have brought about the first explosions.

Lithic content. In addition to pumice and crystals, the ejecta include fragments of old lava torn from the conduit walls and from the upper part of Mount Mazama. These are composed almost entirely of andesite. Bits of dacite may also be found, but ejecta torn from the pre-Mazama basement rocks are extremely rare.

Because of their greater density, the lithic fragments fell close to the source. Had the early explosions blasted away much of the upper part of the volcano, we should expect an abundance of coarse, blocky detritus near the rim of the caldera. No feature of the ejecta is more striking, however, than the paucity of large blocks of old lava dose to the source. Even on Timber Crater, lithic fragments as much as 2 inches across are extremely rare and by far the bulk of such fragments measure less than 1/4 inch in diameter. Still closer to the source, on the west slope of Mount Scott, few of the lithic fragments exceed 3 inches across. Most of the coarser detritus on the rim of the caldera is not of explosive, but of glacial origin.

On the sides of Mount Thielsen, lithic chips larger than 1/2 inch across are practically absent. Chips 1/2 inch across have been found as far as 30 miles from the caldera, at Windigo Pass. On Walker Rim and Skookum Butte, an equal distance northeast of the caldera, where the thickness of the pumice fall is 10 feet and many of the pumice lumps measure 4 1/2 inches across, the largest lithic fragments are less than 1/4 inch in diameter. The maximum size of the lithic fragments in other samples is shown on the histograms, figures 19-22. More than 50 miles from the caldera, chips of old lava exceeding 1/8 inch across are exceptional. The amount of lithic material that fell more than 20 miles from Crater Lake was therefore quite insignificant as compared with the bulk of the pumice and crystals.

In order to determine as far as possible the percentage of rock fragments among the total ejecta of the pumice fall, numerous screen analyses were made in the field, and a few are here recorded in histograms. On Mount Scott and Timber Crater, 4 and 8 miles respectively from the center of Crater Lake, the volume percentage of lithic fragments may reach as high as 12, or fall as low as 4. On Mount Thielsen, the percentage usually varies between 2 and 4, as it does over most of the area south and east of Klamath Marsh. North of the junction of the Willamette and Dalles-California highways, the percentage rarely exceeds 1 and is generally so small that it has been omitted from the histograms, figure 21.

Taking the pumice fall as a whole, the content of old lava fragments averages between 3 and 4 per cent, and in general it is slightly higher in the older part of the fall than in the younger. A similar paucity of lithic detritus has been observed among the pumice deposits of Krakatau, Santorin, and other calderas.⁶ Clearly, the eruption of the pumice fall removed only an insignificant volume of old rock from the summit of Mount Mazama.