Absarokite Banakite Shoshonite Series

In the region of the Yellowstone National Park there are igneous rocks of a peculiar type which are
associated with the normal andesites and basalts of the region, but which differ from them
mineralogically and chemically, and deserves a special classification. The study of these rocks has been
carried on in connection with that of all of the igneous rocks of this region, and the present paper is an
abstract of a chapter prepared for the report of work done in the Yellowstone National Park by the
division of the U. S. Geological Survey under the charge of Mr. Arnold Hague.
The rocks mentioned occur in a number of separate localities within this region, where it is apparent
that their generic relations are with normal basalts and andesites, and in each locality the varieties
having the peculiarities in question are genetically related to one another by differentiation. But all of
these peculiar varieties in the region are not closely related to one another, for they are separate
offshoots from distinct reservoirs of magma, and were probably produced by similar processes of
differentiation. For purposes of systematic description they may be classified together in a series having
certain chemical and mineralogical characteristics. They thus form classes of similar rocks (i. e., like
phases of differentiation), that belong to separate, but similar, families of rocks (i. e., groups or series of
genetically, hence generically, related differentiation products).
For the most part they are basaltic-looking rocks, occurring as lava flows and dikes, and less often as
part of the basic volcanic breccia which constitutes the major portion of the volcanic mountains of the
Absaroka Range. They are quite subordinate in amount to the basalts and andesites. As surficial flows
they are basaltic in character, being dark colored and heavy, with olivine among the phenocrysts in most
cases. They are massive and compact or vesicular. They are porphyritic in some cases, but not noticeably
so in others. They generally exhibit a semi- waxy luster that suggests the presence of nepheline, which
however is not present. The luster is due to the alkali-feldspars in the groundmass. They are often dull
greenish black owing to the serpentinization of olivine. As dikes they are basaltic in some cases, and
phonolitic to trachytic in others, being gray in various shades, and having a somewhat waxy luster, due
to alkali-feldspars. They are porphyritic or not in different cases, and range from aphanitic to
phanerocrystalline. They present quite a range of composition within limits, and form a series of
varieties connected by gradual transitions. They could not be embraced by any one definition and must
be divided into several classes.
The chief characteristics of the most basic class are the presence of abundant phenocrysts of olivine and
augite, and the absence of phenocrysts of feldspar. The groundmass may be anything from a dark glass
to an almost phanerocrystalline, light gray mass. It is oftener aphanitic and dark greenish gray. The
phenocrysts are large and pronounced in many cases, but are quite small in others. Chemically they are
low in silica, from 46 to 52 per cent.; low in alumina, from 9 to 12 per cent.; high in magnesia, from 8 to
13 per cent.; moderately high in alkalies, with potash higher than soda, except in one case. The
molecular ratio of the alkalies to silica is .o8 and .o9. After the crystallization of abundant phenocrysts of
olivine and augite the remainder of the magma, owing to low alumina and relatively high alkalies was so
constituted that alkali-feldspathic minerals might crystallize out, which they did or not according to the
conditions under which solidification took place.

The principal characteristics of the second class are the pres ence of phenocrysts of labradorite, augite
and olivine, in a groundmass that is usually dark greenish gray, with a semi- waxy luster; but which is
sometimes glassy, or phanerocrystal- line. The varieties range from those rich in olivine and augite,
which with decreasing labradorite grade into rocks of the first class, to varieties poor in olivine and
augite. Chemically they range from 50 to 56 per cent. of silica. Alumina is moderate to high, from 17 to
19.7 per cent. Lime and magnesia are moderate to low, the former from 8 to 4.3 per cent., the latter
from 4.4 to 2.5 per cent. Alkalies are moderately high with potash comparatively high for rocks of this
region with like amounts of silica; potash from 3.4 to 4.4 per cent.; soda from 3 to 3.9 per cent. The
molecular ratio of alkalies to silica is .Io and .1 I. After the crystallization of phenocrysts of labradorite,
olivine and augite, the remainder of the magma was rich in alkali-feldspathic material, which shows
itself in the ground- mass according to the circumstances of solidification. Rocks corresponding
chemically to both of these classes occur without megascopic phenocrysts, and in various phases of
crystallization, consequently they differ from them not only in microstructure but in the minerals
developed. They occur as lava flows and as dikes, but no special characteristics can be connected with
either mode of occurrence, except that the more highly crystallized forms are found as dikes. Not all the
dikes, however, are more crystalline than all the lava flows. Rocks of the third class have been found
mostly in dikes. These rocks are highly feldspathic, with smaller amount of ferromagesian minerals,
chiefly biotite with subordinate augite. The phenocrysts are labradorite in a groundmass of alkali-feldspars. Chemically they have from 51 to 61 per cent. of silica ; 16.7 to 19.6 alumina; 3.5 to 6 per cent. of
lime; I to 4 per cent.. of magnesia; 3.8 to 4.5 per cent. of soda, and 4.4 to 5.7 per cent. of potash. The
ratio of alkalies to silica is .13 and .14. Since much of the lime and soda go into the phenocrysts of labradorite, the feldspathic groundmass is rich in potash, and is largely orthoclase. The rocks stand at the end
of the series, where labradorite becomes the prevailing phenocryst, and augite and olivine are more or
less completely replaced by biotite. The division of the series into three parts is for convenience. The
classes will be described in the order just given under the names: Absarokite, Shoshonite and Banakite.

Shoshonite
Shoshonite is a basaltic rock, properly a potassic trachyandesite, composed
of olivine, augite and plagioclase phenocrysts in a groundmass with calcic plagioclase and sanidine
and some dark-colored volcanic glass. Shoshonite gives its name to the shoshonite series and
grades into absarokite with the loss of plagioclase phenocrysts and into bannakite with an increase
in sanidine. Shoshonite was named by Iddings in 1895 for the Shoshone River in Wyoming.
Textural and mineralogical features of potash-rich basaltic rocks of the absarokite-shoshonitebanakite series strongly suggest that most of the large crystals and aggregates are not true
phenocrysts as previously thought but are xenocrysts and microxenoliths, suggesting a hybrid origin
involving assimilation of gabbro by high-temperature syeniticmagma.
Chemical characteristics:
1. Near-saturated in silica;
2. Low iron enrichment;
3. High total alkalies (Na2O + K2O > 5%);
4. High K2O/Na2O;
5. Steep positive slope for K2O versus SiO2 at low SiO2;
6. Enrichment in P, Rb, Sr, Ba, Pb, light rare earth elements;
7. Low TiO2;
8. High but variable Al2O3;
9. High Fe2O3/FeO.
Shoshonitic rocks tend to be associated with calc-alkaline island-arc subduction volcanism, but the
K-rich shoshonites are generally younger and above deeper, steeper parts or the Benioff zone.[4]
Volcanic rocks of the absarokite-shoshonite-banakite series described from Yellowstone Park by
Iddings and the similar ciminite-toscanite series described from western Italy by Washington are
associated with leucite-bearing rocks, potassium-rich trachytes and andesitic rocks. Similar
associations are described from several other regions including Indonesia and the East African Rift.
In the Eolian Arc in the southern Tyrrhenian Sea (between the Eurasian and African tectonic plates),
volcanism has changed between calc-alkaline to high-K calc-alkaline to shoshonitic with the last one
million years, possibly due to the progressive steepening of the Benioff zone, with is inclined at 5060. Late Cretaceous Puerto Rican volcanism is interpreted to have occurred in a
similar tectonic setting.

Absarokite
All of the rocks here classed as absarokites carry abundant phenocrysts of olivine and augite, except two
that are included with them on grounds of chemical identity. They occur in the Absaroka range, and in
other parts of the Yellowstone Park; at the north base of Sepulchre Mountain; upon Mirror Plateau;
within the Crandall volcano; at Signal Point, Yellowstone Lake; at Two Ocean Pass, and in Ishawooa
Canyon, and elsewhere. The chemical composition of five of these rocks is shown by the following
analysis.
Banakite
The most feldspathic rocks of this series, which occur as dikes associated with dikes of shoshonite and
absarokite, are not so numerous, although more of them have been analyzed. They- occur in Crandall
Basin, in Ishawooa Canyon and near the head of Stinkingwater River. A leucite-bearing variety forms a
lava sheet near Beaverdam Creek.
Nomenclature and Chemical profile
The petrographic nomenclature is complex and confusing as it reflects the original definitions given to
similar rock types by several authors (see Joplin, 1968). As a result, some authors even reject specific
names in favour of the more general terms. The nomenclature given in Table 1 is that of Mackenzie and
Chappell (1972) for the shoshonites of the Highlands of New Guinea, to which trachyte and liparite have
been added. This nomenclature applies only for the 'near-saturated' group. For the rarer
'undersaturated' group, Appleton's classification for the leucitic rocks of the Roman Province is used.

Adakite
Adakites are intermediate to felsic volcanic rocks that have geochemical characteristics
of magma that have partially melted from the altered basalt that is subducted belowvolcanic
arcs. Most magmas derived in subduction zones come from the mantle above the subducting plate
when hydrous fluids are released from minerals that break down in the metamorphosed basalt, rise

into the mantle, and initiate partial melting. However, Defant and Drummond recognized that when
young oceanic crust (less than 25 millions years old) is subducted, adakites are typically produced in
the arc. They postulated that when young oceanic crust is subducted it is "warmer" (closer to
the mid-ocean ridge where it formed) than crust that is typically subducted. The warmer crust
enables melting of the metamorphosed subducted basalt rather than the mantle above.
Experimental work by several researchers has verified the geochemical characteristics of "slab
melts" and the contention that melts can form from young and therefore warmer crust in subduction
zones.
The geochemical characteristics Defant and Drummond gave for adakites are:

SiO2 greater than 56 wt %

Al2O3 greater than or equal to 15 wt %

MgO normally less than 3 wt %

Sr greater than 400 ppm

Y less than 18 ppm

Yb less than 1.9 ppm

87

Sr/86Sr usually less than 0.7045

Later Defant and Kepezhinskas reviewed the topic in some detail pointing out that adakites are
found associated with many mineral deposits including gold and copper.
Drummond and Defant noted that Archean trondhjemites (which make up most of the ancient crust
of continents) have similar geochemical characteristics to adakites. They suggested that the entire
Archean crust may have been derived from the partial melting of subducted oceanic crust during the
Archean (> 2.5 billion years ago) because during early earth the temperature of the mantle was
much hotter and more oceanic crust was generated and subducted younger. The proposal has been
controversial and is still being argued among the scientific community. The alternative interpretation
is that the continental crust was derived from the partial melting of lower crustal basalts. The same
idea has also been postulated for the generation of adakites. However, this hypothesis does not
explain the correlation between subducted young crust and adakite eruptions nor the fact that the
lower Yb and Y in adakites suggest that garnet is stable in the source. Garnet forms only under high
pressures within the earth and would not be stable in lower crust below some island arcs that erupt
adakites. See Martin et al. for a more recent summary.
Low magnesium adakites may be representative of relatively pure partial melting of a subducting
basalt, whereas high magnesium adakite or high magnesium andesites may represent melt
contamination with the peridotites of the overlying mantle wedge. Adakites have also been reported
from the continent-continent collision zone beneath Tibet.

Figure 1 Fingerprint for calc-alkaline rocks of Grenadines & Grenada

Figure 2 Trace elements

Figure 3 SiO2 vs. total K2O after Rickwood (1989). The shaded regions denote the divisions between the tholeiite, calcalkaline, high-K calc-alkaline, and shoshonite fields.