Geological Survey Professional Paper 3 The Geology and Petrography of Crater Lake National Park

HYPERSTHENE-ANDESITES.

The andesites are the oldest of the volcanic rocks of Crater Lake and form the great bulk of the exposed rocks. They are all hypersthene bearing, and may therefore be classified as hypersthene-andesites. But in spite of the fact that they may all be placed under this one head, there is a remarkable diversity in their external appearance, inasmuch as they vary in color from black to light gray or light drab and in texture from vitrophyric to holocrystalline. This diversity of texture and color is confined, however, to the groundmass; that is to say, the phenocrysts, although they may show many and varied distinctions, do not appear to be dependent in their variations upon the character of the groundmass to any very great extent. In the following description of these rocks the phenocrysts will be first considered, as they occur in all the andesitic varieties, while the groundmass will be described in connection with the discussion of the different varieties, which depend entirely on the variations of the groundmass.

MINERAL COMPONENTS.

The phenocrysts of these rocks are basic plagioclase, hypersthene, augite, and magnetite. The last-named mineral, to be sure, does not appear as a phenocryst in the hand specimen, but it is always present and belongs in part to the crystallizations of the first generation. To the above should be added olivine and hornblende, which are very seldom present.

PLAGIOCLASE.

As a phenocryst plagioclase occurs usually in stout crystals with sharply developed faces and rather numerous crystal forms, such as have been repeatedly described as characteristic of andesites in general. The most prominently developed form is usually the brachypinacoid (010), the habit of the crystal being thick-tabular parallel to this form. The next most prominent form is the basal pinacoid (001), in addition to which may usually be found a macrodome, probably (101), and the two partial prisms (110) and (110). This habit causes sections parallel to the brachypinacoid to have hexagonal shape and sections cut perpendicular to this face to have very often rectangular or five- or six-sided shapes. The plagioclase crystals usually measure 1 or 2 millimeters in longest diameter. At most the dimensions reach 3 or 4, but not infrequently are considerably less than 1 millimeter. They are nearly always abundant, but macroscopically are conspicuous only in case of the dark-colored specimens with glassy groundmass (2, 3, 17). In the hand specimen these plagioclases are nearly always perfectly fresh and glassy. They can not readily be distinguished from sanidine, owing to apparent absence of cleavage and twinning striations.

In a few cases twinning appears to be entirely absent, even in thin sections, but usually the customary albitic twinning is very pronounced. This may be developed in a few broad bands or in many thin ones that may or may not extend across the crystal. Not infrequently twinning striations at right angles to the first and produced by the application of the pericline law of twinning are to be seen. Crystals thus twinned may be still further complicated by the presence of twinning after the Carlsbad law.

Maximum extinction angles measured to the right and to the left of the trace of the albitic twinning plane on sections cut at right angles to the brachypinacoid indicate in almost every case that these plagioclase phenocrysts are very basic and are to be classed as anorthite. Out of a large number of such measurements the following may be given as indicating fairly average results. In each case the angle given is the largest that was noticed in the specimen referred to.

Maximum extinction angles of plagioclase.

It is not to be inferred from the above that the smaller number of degrees necessarily gives the actual maximum extinction, inasmuch as it is not always possible to find a crystal cut so as to give such angles.

Almost invariably these phenocrysts show a well-developed zonal structure. This is very much more marked in sections parallel to the brachypinacoid than in those cut perpendicular to that face. The general rule in such cases seems to hold true here also, namely, that the extinction angles near the margin of a crystal are less than near the center. In No. 79 the crystal whose extinction angle is given above as 38° for a section cut at right angles to the brachypinacoid gave at the margin an extinction angle of only 30°. This difference of eight degrees for such sections, while perhaps not rare, is larger than ordinary. In very many cases the difference in such sections is not marked. The different extinction zones sometimes gradually shade into each other; sometimes, however, the transition is abrupt, or relatively so. Not infrequently a narrow margin will show a notedly smaller extinction angle than does the rest of the crystal. In still other cases the crystal may be divided into very irregular patches that show either undulous extinction or somewhat different appearing zonal banding in the different parts. A recurrence or alternation of bands that show respectively greater and smaller extinction angles in the same crystal is by no means rare.

As is customarily the case in andesitic rocks, these plagioclase phenocrysts are often characterized by inclosures. These embrace glass, augite, hypersthene, magnetite and apatite, as well as other not well-characterized substances. With the exception of glass these inclosed substances are only occasionally met with, and then any one or more or even all of them may be seen in the same crystal. Even glass, which is so often characteristic of andesitic plagioclase, is by no means always present. In some cases none of the plagioclase phenocrysts contain such inclosures; in other cases some phenocrysts may contain none while others may be filled or perhaps contain only scattered inclosures. These glass inclosures are usually very irregular in form, but they may also be simply spherical, or oval, or polygonal. In color they are sometimes a clear deep brown and are either absolutely free from any crystalline matter (17), or else in inclosures of the same brown color there may be seen microlitic inclusions of augite and magnetite, exactly the same as are to be found inclosed in the brown glass base of the groundmass (16, 6), or, again, a minute gas bubble (6, 26, 17). Such gas bubbles are apparently much more apt to occur in brown glass inclosures in the plagioclase phenocrysts than in the colorless glass inclosures referred to below. Brown glass inclosures are confined mainly to those andesites that have an abundant brown glassy base and that belong to the hyalopilitic type. But exceptions to this rule have been noted (42) where a somewhat paler glass occurs inclosed in plagioclase phenocrysts that are embedded in a hypocrystalline groundmass in which little or no glass base can be readily identified. In this case it is of interest to note that a pale-brown glass of the same color as that of the inclosed glass, and practically free from microlitic or other crystals, occurs in more or less isolated areas formed by the meeting of several phenocrysts of plagioclase or of other minerals. Brown glass inclosures occur nearly always isolated or sparingly scattered through the crystal and not conspicuously parallel to the outlines of the phenocryst. There are, however, a few cases where this brown glass thickly crowds the crystal or a part of it, or occurs in a zone intermediate between a clear center and a clear margin, exactly analogous to what is described below as more characteristic of the colorless glass inclosures in feldspar.

Colorless glass inclosures of very irregular shape and fair size thickly crowding the mass of the feldspar so as to leave the latter in a sort of skeleton form are by no means rare, but they can not be said to be so characteristic as are similar occurrences that will be described in connection with the dacites of Crater Lake. Usually colorless glass inclusions are very minute in size and are densely crowded together so as to give the feldspar a clouded appearance. Such glass inclusions, however, are almost never free from crystalline matter, but contain more or less well-defined material similar to that seen in the groundmass of the hypocrystalline andesites. Owing to this crystalline matter of the glass inclusions the feldspars are rendered not only cloudy but appear dirty, as though fine emery dust had been ground into them. These very minute glass inclusions are not always fresh, but may contain at times ill-defined doubly refracting decomposition matter. The distribution of these beclouding glass inclusions is extremely varied. Sometimes they fill the whole of the crystal mass, although they nearly always leave a clear outer margin. At other times they are aggregated in the central part of the crystal, while still again they may occur in an intermediate zone, leaving both the center and the margin clear. In the last case this clouded zone may be rather broad and not sharply defined either toward the center or toward the margin, or it may be narrow and then more sharply delimited. In general the outer edge of this zone is apt to be more clearly defined than is the inner edge.

These thickly crowded glass inclosures are not to be found in all the plagioclase phenocrysts of the same thin section, nor even in all the phenocrysts of the same general habit. Sometimes they are to be seen in only two or three, while the rest are entirely free from such inclosures. Their occurrence seems to bear some sort of relationship to the character of the groundmass, more particularly to the amount of the glass present in the groundmass. As will appear later, the andesites of Crater Lake will be treated under several heads, based upon the nature of the groundmass. Of these there are three main types, namely: 1, Those with holocrystalline groundmass; 2, those with hypocrystalline groundmass; 3, those with hyalopilitic groundmass. The last-named type contains an abundance of brown glass, the second considerable, but not so evidently developed glass, and the first no glass at all. The relationship of these dense inclosures to the amount of glass may be indicated by the following comparison, which is confined to those phenocrysts that show well-characterized inclosures of thickly crowded glass with a clear margin of feldspar outside: Out of sixteen thin sections of andesite with holocrystalline groundmass, only two, or 12-1/2 per cent, contain such inclosures; out of twenty-one sections of the hypocrystalline type, eleven, or 52-1/2 per cent, contain such inclosures; and out of twenty-one sections of the hyalopilitic type, thirteen, or 62 per cent, contain plagioclase of this description. It would seem, therefore, that the presence of glass in the groundmass has something to do with the crowding of certain of the feldspars with glass inclusions. It should further be stated that plagioclase crystals that have suffered partial resorption, so as to appear in more or less rounded forms, have a stronger tendency to contain such crowded glass inclusions than have those crystals that do not show such resorption. It should also be stated that the above comparison does not take into account the very abundant, but larger and more isolated, inclusions that do not have a marked tendency to occur in an intermediate zone. No such law seems to apply to such inclusions.

The mineral inclosures in plagioclase are not often very abundant. Magnetite occurs in the customary small octahedrons or in grains and apatite in sharp needles or slender prisms. Augite and hypersthene are usually to be seen in roundish grains that have the same color and general appearance as when not inclosed. At times there is to be seen a slightly greenish, finely and irregularly granular mineral, associated with the glass inclusions, that appears to resemble augite more than any other mineral unless it be epidote; but the great freshness of these rocks would seem to exclude the last-named mineral. Another inclosed mineral, that occurs in roundish or oval form and has a slightly yellowish color with rather strong refractive powers, may be zircon, but no sharply formed crystals of zircon were noted.

Resorption phenomena are very common in connection with these plagioclase phenocrysts. They are seen in the rounding of the corners or in the further eating into the feldspars so as to leave them in round grains or in the form of embayments. Such corroded feldspar crystals, as stated above, are peculiarly liable to be filled with glass inclusions to such an extent that there seems to be but a skeleton of feldspar filled with glass. These corroded crystals almost invariably show a clear margin of later growth. The formation of an intermediate zone of glass inclusions is also very strongly developed, and in this case the outlines of the clouded zone are clearly parallel to the outer edge of the newly extended crystal, except where, occasionally, the clear margin has developed new crystal faces. In discussing this well-known phenomenon Professor Rosenbusch says:^a

"Evidently a period of resorption of the already crystallized feldspars is followed by one of rapid addition of new feldspar substance, during which the glass inclosures which are taken up in great quantity on account of the rapid growth must arrange themselves parallel to the deformed periphery. Then came a period of slow growth, with the consequent freedom from interpositions, and the effort to produce regular crystallographic restoration."

---

^aMikroskopische Physiographie, 3d edition, Vol. II, 1896, p. 867.

---

This explanation is plausible and is possibly strengthened by the fact above noted—that the development of this intermediate zone seems in some way dependent on the amount of glass present in the groundmass. It would seem that the period of resorption was followed, in case of those andesites that have developed a holocrystalline groundmass, by a period of slow cooling, which was not favorable to the incorporation of numerous glass inclusions and which continued practically uninterrupted until the whole rock became solidified. Still it does not seem clear just why this phenomenon should be dependent on the present condition of the groundmass, for both the holocrystalline and the partially glassy groundmasses must have been fluid during the period of resorption which is common to both. Another fact often observed in studying these rocks does not seem to fit in with Professor Rosenbusch's explanation. It is this: That the inner line of these intermediate zones is not always as sharp as is the outer one. In fact, it may be quite irregular and not sharply defined, as it should be. In such cases it more closely resembles the gradual encroachment of alteration products upon a clear and unaltered interior. Such feldspar crystals have the appearance of having been honeycombed in the process of resorption, the pores thus produced being filled with the surrounding glass. Such a honeycombing, if possible, would not necessarily advance into the interior of the crystal equally in all directions nor leave always a perfectly sharp line of separation.

The above-mentioned glass inclosures in plagioclase phenocrysts, as well as some of the resorption phenomena and subsequent new growth, are brought out in figs. A, B, C, and D of Pl. XIV (page 76) and in the photomicrographs presented as figs. A, B, and D of Pl. XV (page 80). Figs. A, B, C, and D of Pl. XIV illustrate the distribution of glass inclusions similar to the groundmass also the clear margins of later growth, free from such inclusions. The glass inclusions are rendered black and often opaque through included magnetic dust.

Plate XIV.—CRYSTALS IN ANDESITE, DACITE, AND BASALT.

FIG. A.—Plagioclase in hyalopilitic andesite from Sand Creek. Magnified 30 diameters. Specimen No. 5. The glass inclusions are mainly confined to a narrow zone near the margin. Has undergone partial resorption and secondary enlargement.

FIG. B.—Plagioclase in hyalopilitic andesite from Anna Creek. Magnified 60 diameters. Specimen No. 10. Glass inclusions are uniformly distributed in rather large and irregular patches. Has undergone resorption and slight secondary enlargement.

FIG. C.—Plagioclase in hypocrystalline andesite from water's edge under Llao Rock. Magnified 60 diameters. Specimen No. 31. Glass inclusions uniformly and thickly distributed throughout the original crystal. Has undergone secondary enlargement without resorption.

FIG. D.—Plagioclase in hypocrystalline andesite from Grotto Cove. Magnified 30 diameters. Specimen No. 55. Glass inclusions uniformly and thickly distributed, original crystal partly resorbed, and a secondary enlargement pronounced.

FIG. E.—Andesite at water's edge west of Eagle Cove. Magnified 60 diameters. Specimen No. 11. Shows a phenocryst of augite surrounded in part by a rim of parallel growing hypersthene. See page 82.

FIG. F.—A plagioclase phenocryst. From the vitrophyric dacite of Llao Rock. Magnified 60 diameters. Specimen No. 102. Polarized light with crossed nicols. Shows distinct zonal structure and two directions of cleavage, the basal parallel to the small face near to the figure 3 and the prismatic parallel to the longest side. The section is cut nearly parallel to the brachypinacoid. The intermediate shell 2 is more basic than is the center. See page 101.

FIG. G.—A minute feldspar crystal containing a large inclosure of brown glass with outlines parallel to the sides of the crystal. From the vitrophyric dacite of Llao Rock. Magnified 400 diameters. Specimen No. 101. This glass inclusion contains a single gas bubble which appears like a broad black ring. See page 106.

FIG. H.—Corroded hornblende crystal with resorption rim of augite and magnetite. From the dacite of Grouse Hill. Magnified 200 diameters. Specimen No. 105. See page 111.

FIGS. I, J, K.—Crystals found in the cavities of basalt from the base of Red Cone, No. 156. Figs. I and J are hypersthene and fig. K is pseudobrookite. See pages 146 to 148.

FIG. L.—Olivine phenocryst. From the basalt of Desert Cone. Magnified 65 diameters. Specimen No. 169. Alteration product is like that of fig. M, but alteration is not so far advanced. Contains a plagioclase crystal projecting into the side. See page 153.

FIG. M.—An olivine phenocryst almost completely altered to an opaque mass of which magnetite appears to form the bulk. From the basalt on the inside of the crater of Red Cone. Magnified 110 diameters. Specimen No. 174. See page 155.

Plate XV.—THIN SECTIONS OF ANDESITE.

FIG. A.—Andesite from the south rim of the caldera. Magnified 94 diameters. Specimen No. 1. A photomicrograph in white light. Shows a plagioclase of the first generation, with the center extremely spongiform and crowded with brown glass. The structure thus developed bears a close resemblance to the intergrowth of feldspar and glass in the light-colored granophyric secretions among the dacitic ejectamenta. See page 132.

FIG. B.—Andesite of the hyalopilitic type from east of Sand Creek. Magnified 20 diameters. Specimen No. 15. A photomicrograph in white light. Shows plagioclase crystals of two well-defined types. In the center is a phenocryst of the oldest generation with clear margin and clear center, corroded an one side, and showing an intermediate zone of brown glass inclusions. For descriptions see pages 74 and 84.

FIG. C.—Andesite from the rim south of The Watchman. Magnified 86 diameters. Specimen No. 60. A photomicrograph in white light. Shows a crystal of hypersthene containing small inclosures of glass, with a bubble in each inclosure. Two of these may be seen above the center and to the right and left of the two conspicuous cleavage cracks. See page 81.

FIG. D.—Hyalopilitic andesite from the water's edge west of Eagle Cove. Magnified 20 diameters. Specimen No. 11. A photomicrograph in white light. The section illustrates the occurrence of plagioclase in three generations. First, the oldest, large phenocrysts with glass inclusions, in this case near the margin; second, phenocrysts without glass inclusions and that have mostly rectangular forms; third, small laths of the groundmass feldspar that are mostly too small to be conspicuous in the photograph. See page 77.

In addition to the plagioclase crystals already described, there is very often present a phenocrystic plagioclase that occurs in long to short rectangular sections with square, diagonally extinguishing cross section. These are crystals elongated parallel to the crystallographic axis, and similar, therefore, to the microlitic plagioclases of the groundmass of andesites and basalts. In most cases they are entirely distinct from the groundmass plagioclase, as well as from the above-described phenocrysts. They sometimes form the only plagioclase phenocryst present, but at other times they occur together with the first-described kind and form a second generation of phenocrysts. This type of plagioclase is characteristic of the lavas of Wizard Island (19, 20, 25, 30, 95), but is not confined to that locality. In No. 21, from Sentinel Rock, this type of feldspar is represented, one crystal of which was noticed cutting clearly through a small hypersthene phenocryst. As hypersthene is one of the oldest ingredients in these andesites, this would indicate that these plagioclase crystals can not be considered as part of the groundmass. This is exceptional, however, as in most cases the plagioclase is distinctly younger than hypersthene. They do not appear to be much less basic, if at all, than the plagioclases of the first generation (20 gives maximum extinction angles on sections perpendicular to the brachypinacoid of 30° and 33°). They are usually free or nearly free from inclusions, but they may contain glass inclusions and even be crowded with the same.

From the above it appears that, taking into consideration the plagioclase of the groundmass, this mineral occurs in many of the Crater Lake andesites in three distinct generations. The two generations among the phenocrysts are usually quite sharply defined and readily recognized, but this is not equally true as between the second generation and the third or that occurring in the groundmass. These two are apt to shade into each other. In certain instances it is more than possible that still another generation may be present. Three generations of plagioclase are brought out in fig. D of Pl. XV, which also illustrates a more or less pronounced transition between the plagioclase of the groundmass and that of the second generation. In this case the groundmass feldspars are mostly too small to be readily discernible in the illustration.

The plagioclase of the groundmass varies greatly in size and habit in accordance with the degree of crystallization. In the more glassy andesites it is very apt to assume distinctly microlitic form, so that the microlitic laths may not measure more than 0.005 millimeter in diameter; but more customarily they are considerably larger than this and develop long and short lath forms that gradually pass into the smaller phenocrysts, so that a distinction between the two is not always easily made. Usually these plagioclase laths show extinction angles that are not very much smaller than those of the phenocrysts. For instance, one rock that shows symmetrical extinction angles of 28° and 30° among the phenocrysts contains groundmass laths with extinctions of 24°, 26°, and 27°. The polysynthetic twinning is not always distinguishable, but where this is apparently absent an undulous extinction is very common. In certain of the holocrystalline andesites two generations of feldspar are to be seen in the groundmass—first, an older, lath-shaped variety with developed twinning striae, and, second, an irregularly developed residual feldspar that usually does not show twinning, but which has an undulous extinction. This last-named variety is commonly in irregular, allotriomorphic patches of considerable size, shows simultaneous or nearly simultaneous extinction, and incloses the other minerals of the groundmass, such as plagioclase laths, augite, and magnetite. This residual feldspar will be more fully considered in connection with the description of the holocrystalline andesites. An excellent example is No. 73.

PYROXENE.

Both orthorhombic and monoclinic pyroxenes are characteristic ingredients among the phenocrysts, the former being hypersthene and the latter augite. These two minerals are nearly always abundant, but are never large enough to become prominent in the hand specimen. In general, they are both of about the same size, sometimes one and sometimes the other being the larger. They are not often over 1 millimeter and never more than 2 millimeters in greatest diameter, and from this they may sink to microscopic dimensions. In the hand specimen they can not be distinguished from each other. They have a greenish to brown and brownish-green color and distinctly resinous appearance. They are perfectly fresh, and break without apparent cleavage in roughly conchoidal fractures. Although frequently bunched together, so as to appear larger than they really are, they require a magnifying glass to be clearly seen. Although both of these pyroxenes are almost always present, they vary greatly in their relative abundance. For instance, in a series of six specimens collected along the path descending from the camp ground to the water's level—a vertical distance of about 800 feet—we have hypersthene the more abundant in Nos. 8, 9, and 24, while the reverse holds for Nos. 7, 42, and 44. In general, however, hypersthene is the more abundant (2), although it may become scarce (16). Augite, on the other hand, is rarely more abundant than hypersthene, and is often either scarce (40) or even almost completely wanting, so as to be represented by only one or two roundish grains in a thin section (30, 56).

HYPERSTHENE.

The hypersthene phenocrysts usually occur in well-developed crystals that appear in short to long prismatic habit, seldom in very slender prismatic habit. The forms nearly always present are the brachypinacoids and macropinacoids (100) (010), and the prism (110). The pinacoids are equally developed, and nearly always are more prominent than the prism. This gives to the cross section either a squarish form with truncated corners or an octagonal form. The terminal faces are not so easily determined, but they appear to be a flat pyramid or equally flat domes. In some cases this mineral may occur in more or less irregular grains. This is not apt to be the case when it occurs isolated, but it occurs in this manner when it forms nests, either alone or with augite and magnetite (2). Usually the hypersthene is older than augite, but at times the two appear to have been formed simultaneously.

In not very thin sections hypersthene appears strongly pleochroic, and is then distinctly greenish parallel to the vertical axis and reddish or brownish red at right angles to this axis. These rocks are so fine grained that they require rather thin sections, so that the hypersthenes show much paler colors; but even in the very thin nest sections one can recognize the distinctive pleochroism, as rays vibrating parallel to the vertical axis are always greenish, while those vibrating at right angles to the vertical axis are faintly reddish or yellowish. The three elasticity axes agree with the three crystal axes, as follows: a=a, b=b, c=c. In convergent light sections cut at right angles to the vertical axis give a positive bisectrix and large optical angle, while sections cut parallel to the macropinacoid give a negative bisectrix and a much smaller optical angle. In longitudinal sections the parallel cleavage lines are usually well developed, as is also a cross fracture, but the customary prismatic and pinacoidal cleavages are not often sharply defined in horizontal sections unless the section is unusually thin or the crystals larger than common.

Mineral inclosures in hypersthene are confined to magnetite and apatite, the former being very common and the latter rare. Glass inclusions, however, are very common and characteristic. They are almost always to be seen in the larger and sometimes also in the very smallest crystals, and vary in number from one to twenty or more in case the crystal is unusually large. They are usually roundish or oval, but may be irregular in shape, or they may have polygonal forms resembling that of the crystal. They are usually colorless or have a light-brown shade, but never the deep brown to be seen in many of the feldspars. The color of these glass inclusions does not seem to depend on the color of the glass base nor even upon the presence of such base. Nearly always each of these inclosures may be seen to contain a gas bubble, the size of which does not bear any relationship to the size of the inclosing glass. However, one can often find glass inclosures without gas bubbles in the same crystal in which most of the glass inclusions contain bubbles. The size of such inclusions varies from about 0.01 millimeter to very minute microscopic dimensions. In fig. C of Pl. XV is presented a hypersthene crystal with such inclosures of glass and bubbles.

Hypersthene is almost invariably perfectly fresh. It may at times show a slight trace of serpentinization through the development of a yellowish, fibrous, polarizing substance.

As stated above, hypersthene is to be ranked as one of the oldest crystallizations, but although it occasionally occurs inclosed in the plagioclase phenocrysts it is not always older than this mineral, as the plagioclase sometimes impresses its form on the hypersthene.

In addition to phenocrysts, not a few of the Crater Lake andesites also contain very much smaller hypersthene crystals with a slender lath-like habit. It is not always clear whether these are to be considered as belonging to a later generation and as forming a part of the groundmass or not. In a few cases, at least, of decided glassy varieties (9, 13, 16), this conclusion seems to be justified. In such rocks the groundmass consists of a brown glass containing magnetite grains, a few feldspar microlites, and abundant slender, almost microlitic prisms of augite and hypersthene. The hypersthene is quite uniformly about 0.05 millimeter in length and about one-fifth to one-tenth of that amount in width. The augites are sometimes the same size, but usually not more than half as large. These hypersthenes of the groundmass are sharp, and they frequently show flat terminations. They contain magnetite and also glass inclusions with bubbles. They are mostly very distinct from the hypersthene of the first generation in size and habit and are very much more abundant.

AUGITE.

The phenocrystic augites are, like the hypersthenes, usually sharply crystallized, but they are not so persistently idiomorphic. They not infrequently occur in hypidiomorphic or in granular forms. This, of course, is naturally the case where they form nests either with or without hypersthene. In color they show in thin sections a nearly uniform pale green, the depth of color varying with the thickness of the slide. They rarely show any appreciable pleochroism, and this property, together with the stronger interference colors and the oblique extinction, nearly always suffices to distinguish them from the hypersthenes. Twinning parallel to the orthopinacoid is very common. This is true not only of simple twins, but also of crystals with repeated or polysynthetic twinning.

The inclosures in augite are exactly the same—even to the glass inclusions with bubbles—as are to be found in hypersthene. The crystal form is also very closely analogous to that of the orthorhombic pyroxene. These are the prism with brachypinacoids and macropinacoids and a flat terminal form, presumably a pyramid. Usually, however, the prism and the pinacoids are about equally developed.

Parallel growths of augite and hypersthene are occasionally to be seen (11,171), in which the augite forms more or less irregular shells around the older hypersthene (see fig. E of Pl. XIV, p. 76); but this phenomenon is by no means as common as with the basalts of this region. A more detailed description will be found under the basalts.

As a groundmass constituent augite is very characteristic and is universally present. In proportion to the development of hypocrystalline groundmass, this mineral is to be seen in more and more granular form. In the great majority of specimens, however, this mineral assumes well-defined slender prismatic form. In the hyalopilitic varieties the form is very sharp indeed. The crystals measure two or three or even five times as long as wide. In exceptional cases the length may be proportionately greater still. Usually in sharply developed crystallites the width varies between 0.001 and 0.004 millimeter and the length between 0.01 and 0.03 millimeter. In the holocrystalline varieties the augites are not only granular in shape, but usually much larger. Both microlitic and granular individuals customarily inclose minute magnetite in octahedral and granular forms.

ACCESSORY MINERALS.

Quartz does not appear to form an essential ingredient in these andesites, except as it may form part of the allotriomorphic colorless material to be seen in most of the holocrystalline varieties. With this exception this mineral has been certainly identified in only two cases (46, 79), and in these two cases it does not occur at all evenly distributed, but only in a few irregular grains that may easily be accounted for by secondary or accidental causes.

Tridymite was recognized in only one case, an andesite of a decidedly dacitic type (79), about 1-1/2 miles south of the camp ground.

Hornblende.—A very few yellowish-brown and strongly pleochroic crystals of this mineral, mostly in fragments, are to be seen in specimen No. 44, from near the camp ground, and specimen No. 45, from Cathedral Rock. The larger grains have the black resorption rims of magnetite so common in andesites. In the smaller individuals the resorption is complete and only a black mass of granular magnetite remains (see Pl. XVI, B). Also one small crystal is seen in No. 48, a rock transitional between andesite and dacite. This mineral, which is very common in the dacites of Crater Lake, is otherwise entirely wanting in the andesites.

Olivine occurs very sparingly as an occasional rounded grain or fragment, more particularly among the nests of older crystallized minerals. In nearly all the andesites it is entirely wanting. It is most abundant in No. 49, where it is seen in well-rounded grains that have marked resorption rims of magnetite and, apparently, of augite. In No. 43 there are one or two forms which resemble olivine crystals and which are filled with a yellowish polarizing substance, presumably serpentine.

Magnetite forms an always present but rather sparingly developed ingredient among the minerals of the first generation and occurs frequently inclosed in plagioclase, hypersthene, and augite. In the groundmass it is perhaps more abundant, and then occurs as small octahedral crystals or as grains; also in the andesites with a brown, glassy base, as a very fine powder. It is likewise very abundant as inclosures in the augite microlites of the groundmass.

Hematite appears to occur occasionally either as a reddish stain or as minute brownish-red hexagonal scales (65).