Luhr 1980

Contributions to
Contrib. Mineral. Petrol. 71, 343-372 (1980) Mineralogy and

Petrology
9 by Springer-Verlag 1980
The Colima Volcanic Complex, Mexico

I. Post-Caldera Andesites From Volcfin Colima
James F. Luhr and Ian S.E. Carmichael

Dept. of Geology and Geophysics, University of California, Berkeley, California 94720, USA
Abstract. Volcfin Colima is Mexico's most historically semblage of the olivine-andesite is calculated to equili-
active andesitic composite volcano. It lies 150 km brate at 1,000 bars with 0.8% H20.
north of the Middle America Trench at the western Orthopyroxenes and certain clinopyroxenes in all
end of the Mexican Volcanic Belt, closer to the trench pre-1961 samples are reversely zoned, with relatively
than any other composite volcano in Mexico. Since Mg-rich rims. The most pronounced Mg-rich rims
its earliest reported eruption in 1576, V. Colima has occur in the olivine-andesites and are thought to re-
evolved through three cycles of activity. Each cycle flect pre-eruptive magma mixing, involving a basic,
culminated in a major ashflow eruption, halting activ- olivine + / - clinopyroxene-bearing magma. In addi-
ity for 50 or more years. The last major ashflow tion to their normally zoned pyroxenes, the post-
eruption occurred in 1913. Andesitic block lava erup- 1961, fourth cycle andesites display a number of other
tions in 1961-1962 and 1975-1976 marked the incep- features which distinguish them from earlier post-
tion of activity in a fourth historical cycle which may caldera hornblende-andesites of similar bulk composi-
also terminate with a major ashflow eruption in the tion. These include: (1)higher total crystal contents,
early part of the next century. (2) lower modal hornblende contents, (3) higher calcu-
Major and trace element analyses of whole rock lated pre-eruptive silica activities, and (4) lower calcu-
samples and all constituent phases are presented for lated pre-eruptive water contents. These features are
a suite of nine post-caldera hornblende and olivine- all consistent with the interpretation that the fourth
andesites. The suite includes samples from Colima's cycle andesites were less hydrous prior to eruption.
four major eruptions since 1869, spanning the last The slight Mg-rich pyroxene rims in pre-1961 horn-
two eruptive cycles. Colima's post-caldera andesites blende-andesites may record late-stage, pre-eruptive
are poor in K and other incompatible elements (Ti, increases in magmatic water content, which act to
P, Zn, Rb, Y, Zr, Ba, La, Yb, Hf, Th, and U) as raise magmatic fo2 and Mg/Fe +2 ratios in the melt
may be characteristic of near trench andesites. From and in all crystalline phases. The fourth cycle andesites
the 1913 ashflow eruption through the fourth cycle apparently did not experience a strong, pre-eruptive
andesites, there have been increases in whole rock influx of water, resulting in lower magmatic water
abundances of Si, Ba, and Cs, and decreases in Ti, contents and normally zoned pyroxenes.
Fe, Mg, Ni, Cr, and Sc. Crystal fractionation models
can closely reproduce major element variations in the
post-caldera suite, but systematically fail to predict
sufficient concentrations of the compatible trace ele-
General Geology
ments Cr, Ni, and Zn. Anomalous enrichments of
compatible trace elements in Colima's andesites prob-
Quaternary volcanism in Mexico has been primarily confined to
ably reflect simultaneous crystal fractionation and an irregular, east-west trending belt of large andesitic composite
magma mixing in the subvolcanic system. volcanoes, intervening cinder and lava cones, and rare rhyolitic
Estimated pre-eruptive temperatures range from complexes which are collectively referred to as the Mexican Volca-
940~ ~ C in the hornblende-andesites and nic Belt (Mooser, 1969) (Fig. 1). The main axis of the belt is defined
by a line of large, historically active composite volcanoes including
1,030~ ~ C in the olivine-andesites. Pre-eruptive Volc~m Colima (4,000 m), Popocatepetl (5,452 m), and Pico de
magmatic water contents of 1.0-3.6 wt.% are calcu- Orizaba (5,675 m) (Mooser, 1958). The easternmost, Tuxtla vol-
lated for the hornblende-andesites ; the phenocryst as- canoes, including the historically active Volcfin San Martin, are
0010-7999/80/0071/0343/$06.00
344 J.F. Luhr and I.S.E. Carmichael: The Colima Volcanic Complex, Mexico. I
~A Composite Volcanoes
:'. Cinder Cones in Mexico
Tertiary Rhyolite Tuffs
Volc&n
Ceboruco
Primavera
Rivera
Plate
Rivera
Fracture Zone Vok6n
Colinm
East
Fig. 1. Generalized map of Mexico and Central
Pacific
Rise America indicating major composite volcanoes,
Cocos Plate cinder cones in Mexico, the Tertiary rhyolitic
tufts of the Sierra Madre Occidental, and
principal plate tectonic features of the adjacent
Pacific Ocean basin. Numbered composite
GalSpagos Ridge ~)~o " volcanoes : 1 : Popocatepetl, 2: Iztaccihuatl,
3:Pico de Orizaba, 4:Cofre de Perote, and
Galapagos Islands 5:Volcfin San Martin
constructed of alkalic lavas ranging from picrites to hawaiites and depth earthquakes are generally restricted to the area south of
are compositionally distinct from Mexico's other major volcanoes the Mexican Volcanic Belt, such that few of the major composite
(Thorpe, 1977; Robin and Tournon, 1978). Cinder and lava cones volcanoes appear to be underlain by a seismic zone. The Middle
are largely restricted to the basal plains surrounding composite America Trench trends approximately 110 ~ making an angle of
volcanoes, but several extensive fields exist that are dominated about 15~ with the main axis of the belt. Consequently, the more
by smaller cones. The largest such field is in the western end eastern volcanoes lie at progressively greater distances from the
of the volcanic belt, northeast of Volc~m Colima, and includes trench. This divergence may be related to increased plate conver-
the historically erupted Jorullo (1759-1774; Mooser, 1958) and gence rates in the same direction, or as Mooser (1972) has
Paricutin (1943-1952; Foshag and Gonzalez, 1956; Williams, suggested, to the controlling effect of basement structures on the
1950). locations of volcanic centers.
Throughout the Mexican Volcanic Belt, the locations of major
volcanic centers are fundamentally controlled by basement struc-
tures (Mooser, 1969, 1972). Volc/m Colima lies at the southern
The Colima Volcanic Complex
end of the 90 km long, SSW trending Colima graben, which marks
the western end of the main volcanic belt. At its northern end,
some 50 km SSW of Guadalajara, the Colima graben intersects The Quaternary Colima volcanic complex is roughly circular with
and merges with two other major structural features. The Chapala a radius of approximately 12 km (Fig. 2). It straddles the border
graben, containing Lake Chapala, extends 75 km east from this between the states of Jalisco and Colima, about 150 km northeast
intersection, while a chain of small composite volcanoes trends of the Middle America Trench. Molnar and Sykes (1969) delineated
180 km to the northwest, parallel to the trace of the Tamayo Frac- a crude Benioff Zone about 100 km below the complex, dipping
ture Zone and other transform faults in the adjacent Gulf of Califor- at an angle of 30 ~ to the northeast.
nia. This northwestern segment of the volcanic belt contains abun- The complex consists of two overlapping composite volcanoes.
dant alignments of cinder cones and fault scarps along northwest- The older, extinct cone of Nevado de Colima, centered at
ern trends. Gastil and Jensky (1973) present evidence for major 19~ `` N, 103~ ,' W, marks its highest point at 4,320 m.
right lateral faulting in this zone since the late Cretaceous. The Nevado's peak has been extensively glaciated (Lorenzo, 1961). The
late-Pleistocene Primavera complex of mildly peralkaline rhyolitic younger Volc/m de Colima, Mexico's most historically active vol-
domes and pyroclastics (Mahood, 1977) lies 35 km north of the cano, is centered approximately 5.5 km south of Nevado's peak
junction of these three structures. at 19~ N, 103~ W. The crater dome of Volc/m Colima
Principal plate boundaries of the Pacific Ocean basin adjacent is presently at about 4,000 m elevation. Both Colima volcanoes
to Mexico include the Middle America Trench, the Galapfigos possess large summit calderas as indicated on Fig. 2; each was
Ridge, the East Pacific Rise, and the Rivera Fracture Zone (Fig. 1). partially filled by the later growth of a lava cone.
These separate the oceanic crust into the Cocos and Rivera plates. Immediately north of the Colima complex lies an older,
The Mexican Volcanic Belt presumably developed in response to dissected volcanic center (Fig. 2). Consequently, this area records
the current northeastward subduction of the Cocos Plate into the a progressive trenchward shift in the focus of volcanism by some
Middle America Trench. Molnar and Sykes (1969) estimated rates 20 km over the lifetimes of three composite volcanoes. In the east-
of convergence to be increasing to the southeast along the trench ern portion of the Mexican Volcanic Belt, the historically active
and indicated a poorly defined, northeastward dipping Benioff volcanoes Popcatepetl and Pico de Orizaba each lie at the southern
Zone to depths of 250 km below Mexico. Epicenters of intermediate ends of N-S trending, extinct volcanic chains (White, 1952; Weyl,
J.F. Luhr and I.S.E. Carmichael: The Colima Volcanic Complex, Mexico. I 345
%
19 ~ 50'
/
19040 ,
Fig. 2. Generalized map of the

southern end of the Colima graben,
indicating major fault scarps of the
19~ graben system, the Colima complex
and the older center to the north,
t--" and surrounding cinder cones. The
5
cinder cones are labelled C A B
I I I I I
(calc-alkaline basalt), BA (basaltic
andesite), B (basanite), and M
103 ~ 50" 103 ~ 40" 103 ~ 30' (minette)
1974). Both of these chains also appear to have propagated trench- History and Cyclicity of Eruptions
ward with time. Similar trenchward shifts in the focus of volcanism From Volc~n Colima
occurred through the lives of the paired volcanoes San Pablo and
San Pedro in northern Chile (Francis et al., 1974) and Nekoma
and Bandai in northeastern Japan (Nakamura, 1978). In the Aleu- Volc~n Colima is one of the most historically active volcanoes
tina arc, however, Marsh (1976) noted an opposite, northward in North America. Written accounts of its eruptions date from
migration of volcanism with time. shortly after the Spanish Conquest of Mexico, yielding a record
A 300+ m thick apron of volcanic conglomerates, sands, and of activity over 400 years in length. Compilations of the eruptive
ashes drapes the base of the Colima complex and the older center history of Volcfin Colima were made by Barcena (1887), Sapper
to the north. The upper, gently sloping surface of this clastic apron (1927), Friedlaender (1930), Waltz (1932), and Mooser (1958, 1961).
forms a broad plain surrounding the volcanoes at about 1,500 m These compilations were necessarily based on separate reports vary-
elevation, which becomes coincident with the floor of the Colima ing widely in quality and reliability. Nevertheless, it is apparent
graben as it is followed to the north. In the valley of Rio Tuxpan, that historical activity at Volcfin Colima has evolved in a highly
to the southeast of the complex (Fig. 2), this clastic sequence uncon- periodic manner. Since the earliest available report in 1576, the
formably overlies a greenschist grade metavolcanic basement of volcano has experienced three cycles of eruptive activity. The erup-
unknown age at about 1,000 m elevation. Assuming that the Co- tion of hornblende-andesite block lava flows in 1961-1962 (Mooser
lima complex overlies this basement at the same elevation, its and Maldonando-Koerdell, 1963) and 1975-1976 (Thorpe et al.,
total constructional volume is estimated as 450 km 3. 1977) marked the inception of activity in a fourth historical cycle.
Along its length, the floor of the Colima graben is alternately Figure 3 is a plot of all reported eruptions of Volc~m Colima 1
occupied by clastic aprons, playa lakes, and recent, small volume designated as: (1) unspecified, (2) block lava, (3) airfall ash and
volcanic centers. At the southern end of the graben, a dozen or cinder, and (4) maximal, Pelean-type ashflow eruptions. Those
more recent cinder cones have erupted through the clastic aprons eruptions whose products were sampled for study are briefly de-
surrounding Nevado de Colima and the older volcano to the north scribed in the next section. The last three eruptive cycles at Volcfin
at 900 m-l,700 m elevation (Fig. 2). Many of these cones have Colima have culminated in major ashflow eruptions. Following
well preserved craters and may be only tens of thousands of years the cycle-ending ashflow eruption of 1913, the 350 m diam. summit
old. Nine of the cinder cones have been sampled for study. Scoria crater was cleared to a depth greater than 60 m (Mooser and
from seven of them form a transitional series from potassic basa- Maldonando-Koerdell, 1959). At such times, the crater of V. Co-
nites to minettes. The occurrence of basanites and minettes at lima is probably very similar to the present, open craters of Popo-
the southern end of the Colima graben is consistent with the catepetl and Pico de Orizaba (Mooser, 1958). The next eruptive
general association of alkalic volcanic centers and areas of exten- cycle then begins with a slow, discontinuous, piston-like ascent
sion. To our knowledge, the minettes represent the youngest and of the lava dome in the open crater. During phases of lava dome
freshest known samples of iamprophyric magmas. The petrology
of these cinder cones is considered in Luhr and Carmichael (i979) 1 A detailed history of eruptions is available from the authors
and in Part II of this paper. on request
ERUPTIVE PERIODICITY AT VOLCAN COLIMA
.T
D UNSPECIFIEDERUPTION
e BLOCK LAVA ERUPTION
9 AIRFALL ASH AND CINDER ERUPTION
MAXIMAL ASHFLOW ERUPTION
CYCLE 1 ~]4 CYCLE 2 :!.CYCLE 3=!. CYCLE 4
CRATER DOME
ASCENT
1500 1600 1700 1800 1900 2000
ERUPTION DATE
Volcan de Colima
ERUPTIVE CYCLE I 2 3 4
PRECEDING ASHFLOW -- I•! 1818 1913

• 1975-76 hbd. andesite
ERUPTION DATE block lava
INITIAL 1576 1749 1869 1961 ~ 1961-62 hbd. andesite
ERUPTION DATE block lava
FINAL ASHFLOW 1611 1818 1913 1913 hbd, andesite
ERUPTION DATE
ashflow deposit
DURATION OF 35 69 44
ACTIVITY lyrsJ 1869 hbd andesite
block lava
DURATION OF -- 138 51 48 1 km
INACTIVITY {yrs.] prehistoric ol. andesite
DURATION OF ~ 207 95 block lava
CYCLE [yrs.] caldera wall
99 sample locations
Fig. 3. Plot of historical eruptions of Volcdn Colima versus time and numbers
and a list of data used in defining historical eruptive cycles. Sources
of data: Barcena (1887), Sapper (1927), Friedlaender (1930), Waitz
(1932), Mooser (1958, 1961), Mooser and Maldonando-Koerdell Fig. 4. Generalized m a p of the s u m m i t of Volcdn Colima indicating
(1963), and Thorpe et al. (1977) sampled, post-caldera flows
ascent, which have lasted 50 140 years, eruptive activity is minimal been approximately 0.002km3/yr. Extrapolation of this rate
to nonexistent (Fig. 3). At the termination of the dormancy phase, implies an age for the post-caldera cone and a m i n i m u m age for
the dome is generally in the vicinity of the crater rim. Eruptive the caldera of 2,500 years.
activity resumes in the form of pyroclastic or block lava eruptions. Nine samples were collected from seven of the most recent,
These phase initiating eruptions occurred in 1576, 1749, 1869, and post-caldera andesites at VolcS.n Colima. Sample locations are indi-
1961. Once begun, activity continues intermittently for 35-70 years cated on Fig. 4 and described below in chronological order of
and terminates with a maximal, ashflow dominated eruption. Cy- eruption. A brief history of each sampled eruption is given when
cle-ending ashflow eruptions occurred in 1611, 1818, and 1913. possible.
Following each culminating eruption, the crater dome of the next Col 11 and 13. Several large, prehistoric flows of olivine-
eruptive cycle begins its discontinuous ascent to the rim. andesite form the floor in the northwestern part of the caldera.
The eruptions of 1961-1962 and 1975-1976 marked the begin- One of the stratigraphically oldest lobes was sampled as Col 13,
ning of a fourth historical phase of activity at Volcfin Colima. and one of the youngest as Col 11. These olivine-andesites are
By analogy with eruptive behavior exhibited in earlier cycles, it considered to be the oldest, accessible post-caldera flows at Volcfin
is expected that during the next 20-50 years, pyroclastic activity Colima.
will be c o m m o n with lesser likelihood of further lava eruptions. Col 7. - Col 7 was sampled in April, 1975, from a prominent
The fourth cycle should end sometime in the early part of the peak on the eastern crater rim which was subsequently buried
next century with one or more large, crater-clearing ashflow erup- by the flows of 1975-1976. The peak is visible as the high point
tions. Calculations in this paper, however, point to a significant on the crater rim in the 1913 photograph of Waitz (1915) and
reduction in magmatic water content during the current eruptive in the 1961 photograph of Mooser and Maldonando-Koerdell
cycle, which may lead to more subdued, less explosive behavior (1963). The exact age of Col 7 is unknown, but it can confidently
of the volcano. be placed prior to the 1869 flows in order of eruption.
Col 18 and 2. Following the maximal ashflow eruption of
February 5, 1818 (Mooser, 1961) which is interpreted as having
ended the second eruptive cycle, no activity occurred at Volc~m
Post-Caldera Lavas and Scoria
Colima for 51 years as a lava dome slowly rose in the open crater.
In June of 1869, the dome was still below the crater rim, but
The active cone of Volc~n Colima (4,000 m) rises 1-1.3 k m above the rising m a g m a found an opening in the northeastern side of
the westward sloping, infilled floor of the caldera, which itself the main cone and the parasitic cone of Volcancito was formed
is about 1.4 k m above the surrounding Colima graben floor. The about 1 km from the axis of the main crater and 100 m below
cone has overflowed and obliterated the caldera walls on all but the its rim. Approximately 0.19 km 3 of block lava was erupted from
northern and northwestern sides (Fig. 4). Initial caldera diameter Volcancito, filling the northeastern portion of the caldera and over-
is estimated as 4.5 km; the volume of the active, post-caldera cone flowing and obliterating the eastern caldera wall (Sartorius, 1871;
is approximately 5 km 3. Based on reported eruptions over the last Waitz, 1906; Mooser, 1961). Col 2 was collected from the northern-
200 years, the average growth rate for the intracaldera cone has most flow front. Sometime after the Volcancito related flows were
erupted, the main crater overflowed its rim and a double-leveed, time a lava dome slowly rose to the crater rim. The dome stagnated
bilobate flow with an estimated volume of 0.08 k m 3 was emplaced in about 1935 as a fiat lava plug 60 m below the rim. In May
on the northwestern face of the cone. Sample t8 was taken from of 1957, fumarolic and solfateric activity resumed from cracks
its flow front. The 1869 flows are characterized by a b u n d a n t large in the dome as it began arching and rising to the rim accompanied
hornblende crystals (Table 6). by small earth tremors (Mooser and Maldonando-Koerdell, 1959).
Col 1 5 . - After the major 1869 eruption, Volcfin Colima exper- The dome spilled over a low spot in the northern rim in mid-1961.
ienced occassional minor and major pyroclastic eruptions until F r o m then until late-1962, a short block lava flow of approximately
1909 (Arreola, 1903; Sperry, 1903; Mooser, 1961), when it entered 0.02 km 3 volume descended the northern face of the cone to the
a four year period of inactivity. On January 20, 1913, a tremendous caldera floor, a distance of about 300 m (Mooser and Maldonando-
ashflow eruption began (Waitz, 1915), closing the third, historical Koerdell, 1963). The 1961 1962 activity is interpreted as having
eruptive cycle. Ash clouds from this eruption traveled to the north- initiated the eruptive phase of the fourth historical cycle. Col 17
east, dropping approximately 0.5 m m of ash of Guadalajara, about was sampled from the flow front.
140 k m distant. M o s t of the erupted material was deposited during Col 6. Col 6 was sampled in April, 1975 from the 13 year
the first day of activity by ashflows which moved down canyons old crater dome prior to the 1975-1976 eruption. It is assigned
on all sides of the volcano, extending up to 8 k m from the crater. an eruptive date of 1962 in this report. Stoiber (1967) investigated
The 1913 activity had a dramatic impact on the morphology of fumaroles and fumarolic sublimates from the crater dome of Vol-
the upper cone, impressively illustrated by photographs ac- cS,n Colima in February, 1966. Sulfur, halite, sylvite, thenardite,
companying Waitz's (1915) report. Col 15 is a scoria sample from soda alum, anhydrous sodium a l u m i n u m sulfate, and yellow iron
near the base of the 7 m thick ashflow deposit forming the flat chlorides were identified a m o n g the sublimates.
caldera floor to the north of the cone. Partial wet chemical and Col 9. - Between December, 1975 and June, 1976, the crater
X-ray fluorescence analysis of the dense scoria now filling the dome of Volcfin Colima overflowed the northeastern and eastern
1869 vent of Volcancito demonstrates it to be compositionally crater rims. Three small lobes of blocky andesite flowed northward
identical to the Col 15 scoria, while distinguishable from every toward the caldera floor and two other long tongues moved down
other analyzed sample. Since the only major, post-1869 ashflow major canyons on the southeast side, reaching almost to the base
eruption of Volc~m Colima occurred in 1913, we are confident of the volcano. Total eruptive volume is estimated as 0.12 k m 3
in interpreting Col 15 as a product of the cycle-ending, 1913 activ- The downslope movement of these lavas was accompanied by fre-
ity. quent block and ash-flows from the oversteepened flow fronts
Col 17. - Following the tremendous ashflow eruption of 1913, (Thorpe et al., 1977). Col 9 was sampled from the front of the
the volcano was essentially d o r m a n t for 48 years, during which largest, northern flow lobe in April, 1976.
Table 1. Whole rock wet chemical analyses
Col 9 6 17 15 18 2 7 11 13
SiO 2 61.02 60.41 60.41 57.57 61.09 60.84 61.61 56.55 56.46
TiO 2 0.64 0.67 0.72 0.79 0.61 0.62 0.60 0.78 0.77
A120 3 17.72 18.08 17.17 17.42 17.73 17.80 17.82 16.63 16.67
Fe203 1.77 2.08 1.89 2.64 1.96 2.16 2. I 0 1.46 2.11
FeO 3.46 3.25 3.92 3.74 3.18 3.10 2.96 5.23 4.56
MnO 0.i0 0.10 0.12 0.12 0.10 0.11 0.10 0.11 0.12
MgO 2.76 2.80 3.22 4.14 2.60 2.74 2.54 5.94 5.98
CaO 5.92 6.31 5.89 7.02 5.84 5.88 5.70 7.75 7.65
Na20 4.70 4.78 4.67 4.40 4.78 4.76 4.77 4.03 4.02
K20 1.40 1.31 1.40 1.16 1.41 1.46 1.43 1.06 1.06
P205 0.21 0.20 0.22 0.19 0.19 0.22 0.20 0.20 0.17
H20 + 0.06 0.09 0.13 0.49 0.07 0.10 0.10 0.15 0.18
H20 0.10 0.06 0.07 0.11 0.07 0.04 0.05 0.05 0.05
Total 99.86 100.14 99.83 99.79 99.63 99.83 99.98 99.94 99.80
CIPW n o r m (tool%)
qz 10.48 9.43 9.49 6.69 10.70 10.29 11.54 3.34 4.01
or 8.25 7.69 8.26. 6.86 8.33 8.60 8.42 6.21 6.23
ab 42.10 42.66 41.87 39.58 42.90 42.63 42.68 35.91 35.89
an 23.07 23.86 21.72 24.40 22.75 22.83 22.91 23.98 24.17
di 3.89 4.76 4.82 7.40 3.98 3.86 3.18 10.31 9.97
hy 9.04 8.09 10.40 10.80 8.05 8.21 7.84 17.24 16.11
mt 1.85 2.16 1.97 2.76 2.05 2.25 2.19 1.51 2.19
il 0.89 0.93 1.00 1.10 0.85 0.86 0.83 1.08 1.07
ap 0.44 0.42 0.46 0.40 0.40 0.46 0.42 0.42 0.35
Analyses by I.S.E. Carmichael

Alkalis by J. Hampel, flame photometer
Whole Rock and Groundmass Chemical Compositions margins, with continental andesites commonly containing higher
concentrations of Si and incompatible elements (Jakes and White,
1972). In m a n y volcanic arcs, the concentrations of K 2 0 and re-
Wet chemically determined major element analyses and CIPW
lated elements increase with distance from the trench (Kuno, 1959;
n o r m s of the nine whole rock samples are listed in Table 1 in
Rittmann, 1962; Sugimura, 1968; Hatherton and Dickenson, 1969;
order of increasing sample age. The alkali-lime index for Colima's
Katsui et al., 1978). Volc~n Colima lies 150 k m from the Middle
post caldera suite, constructed from a Harker variation diagram,
America Trench, closer than any other composite volcano in Mex-
is 60.6, at the calcic end of the calc-alkalic class (Peacock, 1931).
ico. Colima's investigated post-caldera andesites have relatively
This is simply a reflection of the high calcium/low alkali nature
low concentrations of K, Ti, P, and incompatible trace elements
of the suite.
(Zn, Rb, Y, Zr, Ba, La, Yb, Hf, Th, and U). Colima's andesites
19 trace elements and 3 major elements determined by neutron
are contrasted with the incompatible element enriched post-caldera
activation analysis for 8 of the samples are listed in Table 2. 6
andesites from the Mexican Volcano Ceboruco (250 km from the
trace elements determined by X-ray fluorescence are tabulated for
trench) in Luhr and Nelson (1979).
each sample in Table 3. Wet chemical and neutron activation anal-
yses of the groundmass phase in seven of the samples are given
in Tables 4 and Table 5 respectively. As is characteristic of m a n y
Temporal Variations in Magma Composition
calc-alkaline andesites (Taylor, 1969), Colima's post-caldera ande-
sites do not display significant Eu anomalies (Table 3). Moorbath
et al. (1978) determined SrS7/Sr 86 ratios for 8 andesites and basaltic A number of recent studies have attempted to determine the nature
andesites from the Colima volcanoes, ranging from 0.7036-0.7040. and extent of variations in m a g m a compositions with time at spe-
Calc-alkaline andesites vary widely in major element, trace cific andesitic volcanoes (Condie and Swenson, 1973 ; Katsui et al.,
element, and isotopic compositions throughout the world. Classifi- 1975 ; Ando, 1975; Rose et al., 1977 ; Newhali, 1978 ; Fairbrothers
cation schemes for orogenic andesites generally rely on KzO as et al., 1978; Stern, 1979; Nelson, 1979; Sakuyama, 1979). While
a principal discriminant (Taylor, 1969), although a n u m b e r of other each volcano has unique aspects to its chemical evolution, most
incompatible elements m a y vary sympathetically with K20. High studies have documented periodic alternations between basic, inter-
and low K andesites occur in both island arcs and on continental mediate, and occasionally silicic m a g m a s superimposed on an over-
Table 2. Neutron activation analyses of whole rock samples a
Coi 9 6 17 15 2 7 11 13
Na% 3.66 3.56 3.64 3.35 3.67 3.68 3.13 3.15

Sc 12.7 13.6 14.9 18.8 12.0 12.1 23.9 23.4
Ti% 0.31 0.38 0.36 0.40 0.32 0.31 0.41 0.37
V 150 140 196 179 169 130 229 231
Cr 25 32 35 75 34 17 210 210
Mn 811 798 900 913 833 778 991 958
Fe% 3.86 4.02 4.44 4.74 3.90 3.83 5.15 5.04
Ni 10 25 26 51 32 21 80 73
Zn 70 66 73 66 63 62 77 66
Cs 1.0 0.7 0.8 0.4 0.7 0.6 0.4 0.6
Ba 510 463 497 413 494 539 391 369
La 12.4 11.7 12.8 10.4 12.7 12.6 9.5 9.3
Ce 27.1 26.4 28.4 24.9 28.3 24.5 22.7 21.7
Nd 12 12 15 15 15 13 14 16
Sm 2.93 2.90 3.17 2.97 2.97 2.73 2.94 2.86
Eu 0.98 0.99 1.01 1.01 0.99 0.90 0.99 1.00
Tb 0.43 0.37 0.47 0.45 0.42 0.38 0.41 0.44
Dy 3.1 2.7 3.1 2.9 2.9 2.6 2.7 2.8
Yb 1.83 1.81 1.96 1.77 1.69 1.68 1.72 1.73
Lu 0.25 0.09 0.t6 0.25 0.15 0.14 0.16 0.15
Hf 3.5 3.4 3.8 3.1 3.5 3.3 2.9 2.7
Th 2.01 1.92 2.09 1.52 1.84 1.98 1.27 1.10
U 0.78 0.81 0.75 0.56 0.74 0.80 0.54 0.53
Eu/Eu *b 1.01 1.06 0.95 1.01 1.01 1.00 1.02 1.04

ZREE's 61.0 59.0 66.1 59.7 65.1 58.5 55.1 56.0
Th/U 2.58 2.37 2.79 2.71 2.49 2.48 2.35 2.08
In ppm, except Na, Ti, and Fe

All neutron activation analyses by F. Asaro and H. Bowman, Lawrence Berkeley Laboratory (Perlman and Asaro, 1969)
b Eu* is interpolated between Sm and Tb
Average counting uncertainties of one standard deviation are equal to the following percentages of the amounts present: N a (1%),
Sc (1%), Ti (5%), V (15%), Cr (3%), M n (1%), Fe (2%), Ni (10%), Zn (10%), Cs (10%), Ba (5%), La (20%), Ce (10%), N d
(15%), Sm (2%), Eu (2%), Tb (15%), Dy (20%), Yb (5%), Lu (15%), H f (10%), Th (10%), U (5%)
Table 3. X R F analyses of whole rock samples Table 5. N e u t r o n activation analyses of groundmass samples ~
Col 9 6 17 15 18 Col 9G 6G 17G 15G 2G 7G llG
Cu 27 35 30 23 18 Na% 3.94 4.07 4.16 3.94 4.16 3.90 3.90

Ga 19 19 17 17 18 Sc 9.3 9.0 9.0 11.3 6.9 9.5 16.1
Rb 19 20 21 16 21 Ti% 0.40 0.35 0.34 0.38 0.25 0.33 0.54
Sr 568 593 531 584 609 V 110 93 62 118 55 99 301
Y 17 17 18 20 17 Cr 8 4 4 6 3 4 29
Zr 152 154 159 131 134 Mn 645 570 535 732 670 613 810
Fe% 2.93 2.55 2.59 3.63 2.63 2.97 4.86
K/Rb 595 559 563 615 557 Ni 11 <10 10 16 12 <10 27
Rb/Sr 0.034 0.033 0.039 0.027 0.034 Zn 65 64 58 65 68 62 72
Cs 1.8 1.6 1.5 1.0 1.3 1.3 1.1
Col 2 7 11 13 Ba 690 765 810 595 728 705 624
La 20.0 20.0 19.6 16.0 16.6 15.7 15.7
Cn 21 19 35 30 Ce 43.2 43.0 45.7 35.3 38.3 35.7 38.2
Ga 19 20 18 21 Nd 22 22 23 19 20 19 23
Rb 22 20 13 15 Sm 4.33 4.53 4.62 3.84 3.66 3.67 4.40
Sr 625 602 568 571 Eu 1.07 1.13 1.16 1.22 1.04 1.00 1.38
Y 15 17 20 16 Tb 0.59 0.60 0.64 0.48 0.46 0.53 0.62
Zr 152 139 118 115 Dy 4.4 4.3 4.5 3.7 3.3 3.3 4.2
Yb 2.71 2.91 2.94 2.25 2.13 2.24 2.45
K/Rb 550 598 677 603 Lu 0.35 0.41 0.40 0.32 0.30 0.32 0.36
Rb/Sr 0.035 0.033 0.023 0.026 Hf 6.2 6.4 6.7 4.4 5.0 4.9 4.7
Th 3.62 3.84 3.86 2.38 2.92 3.17 2.40
Counting uncertainties of one standard deviation are: Sr (5%), U 1.27 1.38 1.42 0.91 1.09 1.15 0.88
Rb (13%), Y (22%), Zr (7%), Cu (15%), and Ga (10%) of the
a m o u n t s present In ppm, except Na, Ti and Fe; uncertainties as in Table 2
Table 4. Wet chemical analyses of groundmass all trend toward more silicic compositions. Such compositional
trends are generally disrupted by caldera forming eruptive events,
Col 9G 6G 17G 15G 2G 7G llG which usually involve silicic magmas. The c o m m o n occurrence
of summit calderas at andesitic composite volcanoes argues
SiO2 67.65 68.70 68.82 61.90 66.80 66.70 60.45 strongly for the presence of relatively large, near surface m a g m a
TiO2 0.70 0.73 0.70 0.76 0.61 0.61 1.11 chambers below or within the volcanic pile. Trends toward more
A1203 15.37 15.10 15.10 16.99 16.31 16.08 16.91 evolved eruptive compositions with time imply that some mecha-
Fe203 1.37 1.26 1.25 2.17 1.06 1.66 1.80 nism such as crystal fractionation is operating within these
FeO 2.64 2.29 2.23 2.72 2.39 2.34 4.55 chambers. Wide compositional alternations in erupted m a g m a
MnO 0.06 0.07 0.05 0.08 0.08 0.07 0.11 types, however, demonstrate that the volcanoes periodically receive
MgO 1.13 0.85 0.75 2.80 1.20 1.21 2.26 pulses of relatively basic m a g m a from depth. Realistic chemical
CaO 2.99 2.77 2.77 4.68 3.58 3.59 5.28 models for temporal variations in eruptive composition at andesitic
Na20 5.18 5.41 5.42 5.15 5.39 5.24 5.30 volcanoes, therefore, must be capable of simultaneously treating
K20 2.53 2.51 2.51 1.77 2.19 2.18 1.74 crystal fractionation, m a g m a mixing, and other processes (O'Hara,
P205 0.22 0.21 0.22 0.28 0.26 0.22 0.22 1977).
H20 + - - - 0.71 - - - Selected major and trace element concentrations for the post-
H20- - . . . . . caldera andesites of Volcfin Colima are plotted versus eruption
date in Figs. 5 and 6. Arbitrary, prehistoric eruption dates are
Total 99.84 99.90 99.82 100.01 99.87 99.90 99.73 used for Col 7, 11, and 13. Temporal variations in the investigated
suite can be viewed on three time scales, covering: (i) all of post-
CIPW norms (mol%) caldera time, (2) a single eruptive cycle, and (3) a single eruption.
qz 18.18 19.14 19.41 10.67 16.24 17.20 7.53 The olivine andesites (Col 11 and 13) are both the oldest and
or 14.95 14.81 14.83 10.44 12.88 12.86 10.26 the most basic of the investigated samples. In a general way, then,
ab 46.53 48.53 48.66 46.17 48.19 46.97 47.50 compositions have become more evolved through the life of the
an 11.22 9.50 9.46 17.99 13.78 13.89 17.18 post-caldera cone, but reversals in this trend have also occurred
di 1.75 2.29 2.28 2.70 1.77 1.96 6.03 (i.e., Col 15).
hy 4.50 2.96 2.62 8.14 4.64 4.08 7.63 Correlation between temporal variations in m a g m a composi-
mt 1.43 1.32 1.31 2.27 1.10 1.73 1.88 tion and the historical eruptive cycles of VolcS_n Colima is made
il 0.98 1.02 0.98 1.06 0.85 0.85 1,54 difficult by a lack of adequate sample control. No samples can
ap 0.46 0.44 0.46 0.58 0.54 0.46 0.46 be positively correlated with known eruptions prior to 1869, and
the numerous pyroclastic eruptions between 1869 and 1913 (Fig. 3)
500-
p~n 4 0 0 -
~// - .... i.. ~ I
62- -200
//e .......... IP,, \/S'lO 2
60-
\ -100 ppm
wt%58 - ii I \\\ 8O
d
/ 9", . . . . . . . . . Cr- -0
60- \\
56-
\, / /
40-
~'.. "~n. . . . . . . . -uk" /-'~LmTJO 2 -.6 \ _ _ .-dK"
18- 20-
wt%17 [ /
/
A "" iI
-20
-8 ~\ //t .
-- ppm
~.'. _ J ~ C a O -6
--IO
\ . . . . . . . . . ~ /~v 2.0
-4wt.% ~___~ ~___ ~ Y b :
~ \\ /:,~--..~ -
Na2u 1.5-
"~- ~,'/" - ' - ' M g O
-2 ppm 1.0- Cs
1.5
wt%
~,- ,b..~.Ij..~K 20
.5-
1 /
1.0
/N~ I I O-
1800 19100 ' 20100
ERUPTION DATE 18~@ i 1900 t 2000
ERUPTION DATE
II
13 11
I
7
II
2 18
I
15
Ill
17 6 9 II I II I III
SAMPLE NUMBER 13 I1 7 2 18 15 17 6 9
SAMPLE NUMBER
- - '1' ,I, -- -- ,1' '1'
CYCLE 3 CYCLE 4 CYCLE 3 CYCLE 4 - - -
I p
DOME DOME DOME I DOME I
ASCENT ASCENT ASCENT ASCENT
Fig. 5. Selected major element analyses of whole rock samples Fig. 6. Selected trace element analyses of whole rock samples plot-
plotted versus eruption date. Dashed lines indicate inadequate sam- ted versus eruption date. Dashed lines indicate inadequate sample
ple control between points control between points. Error bars reflect uncertainties listed in
Table 2
are not represented in the sample suite. A reliable, nearly complete have been sampled and analyzed in duplicate. Col 11 and 13 are
record is only available from 1913 to the present. Over that time, from different lobes of the olivine-andesites which filled the north-
from the final ashflow eruption of the third eruptive cycle (Col 15) western portion of the caldera, and Col 18 and 2 are from the
through the fourth cycle andesites (Col 17, 6, and 9), there have two different hornblende-andesite flow lobes erupted in 1869. De-
been fairly continuous increases in SiO2, Ba, and Cs, and decreases spite rather large differences in compositions and zoning patterns
in TiO2, FeO r, MgO, Ni, Cr, and Sc. Deviations from this overall of mineral phases in the sample pairs, whole rock major and trace
trend may reflect local crystal accumulation or separation as element compositions are nearly identical (Tables 1-3). In two
suggested by the relatively high Na, Ca, A1, and Sr, and low tested cases, then, there appears to be little variation in chemical
Ti, Fe, and Mg contents of Col 6. Major and trace element crystal composition through the course of lava eruptions at Volcfin Co-
fractionation models for the generation of compositional trends lima. Fairbrothers et al. (1978) noted a similar degree of homogen-
from Col 15 to Col 9 are presented in a later section. Because eity within the 0.01 km 3 andesitic flow erupted from Boqueron
of the limited data base available, no firm conclusions can be Volcano, E1 Salvador, in 1917. Nelson (1979) analyzed four samples
drawn concerning possible controls on eruptive periodicity imposed from the 1.1 km 3 dacitic lava erupted from Volc~n Ceboruco, Mex-
by magma compositions. The major ashflow eruption of 1913, ico, in 1870. He noted a slight decrease in SiO 2 through the eruption
which terminated the third historical eruptive cycle, involved one (0.6 wt.%), but no systematic variations of other major or trace
of the most basic of the post-caldera andesites. It may be speculated elements.
that through the active phase of an eruptive cycle, magma composi-
tions become progressively more evolved with time (Col 17 to
Col 9), but that the cycle ending ashflow activity (Col 15) abruptly
reverses this trend, halting activity for 50 or more years. Analysis
Mineralogy and Petrology
of future eruptive products from Volc/m Colima, particularly of
the anticipated ashflow eruption to end the present eruptive cycle, The nine samples considered in this study are readily divisible
will allow this hypothesis to be tested. into hornblende-andesites (Col 9, 6, 17, 15, 18, 2, and 7) and oliv-
Compositional variations with time during a single eruption ine-andesites (Col 11 and 13) based on the mutually exclusive pres-
can be radically different from variations observed through a se- ence of either mineral as a stable, crystallizing phase. In addition,
quence of eruptions over a period of hundreds to thousands of each sample contains phenocrysts and microphenocrysts of plagio-
years (Katsui et al., 1975). Two of Colima's major lava eruptions clase, orthopyroxene, clinopyroxene, and titanomagnetite, and very
Table 6. Modal analyses (vol.%) ( > 700 points counted)
Col Eruption plag cpx opx oxides hbd oliv grndm sum
date xlls
9ph 1975 30.4 1.8 1.9 - 0.8 tr 52.8 47.2

9mp 7.7 1.2 1.2 2.1 0.1 --
6ph 1962 30.0 3.2 2.7 0.I 0.6 tr 49.5 50.5
Stop 8.I 0.8 2.6 2.4 -- --
17ph 1961 31.8 2.0 3.8 0.1 1.0 tr 48.4 51.6
17mp 8.4 0.7 2.3 1.5 -- --
15ph 1913 14.7 0.4 2.6 - 3.6 tr 70.4 29.6
15mp 3.6 1.8 1.8 0.8 0.4
18ph 1869? 13.9 0.5 0.6 0.4 4.9 tr 64.2 35.8
18rap 13.8 0.1 1.1 0.4 0.1 --
2ph 1869 23.1 1.6 0.9 0.2 3.5 tr 62.0 38.0
2mp 4.8 0.6 1.5 1.7 0.1 -
7ph ? 22.2 1.8 2.9 -- 1.5 tr 61.3 38.7
7mp 5.8 1.3 1.2 1.8 0.2 --
1 lph ? 21.9 4.5 3.2 -- -- 1.3 60.0 40.0
llmp 2.9 2.5 2.1 -- 1.6
13ph ? 23.7 2.0 2.5 -- -- 2.0 54.9 45.1
13rap 7.8 3.9 2.0 0.7 -- 0.5
ph: phenocrysts ( > 0.3 m m ) ; mp : microphenocrysts ( > 0.03 m m )
rare microphenocrysts of hematite-ilmenite set in a microcrystalline Table 7. Brown glass inclusions are most a b u n d a n t in the oldest,
to glassy groundmass. All samples have glomeroporphyritic tex- most basic, olivine-andesites, where heavily riddled cores are sur-
tures. Following the grain size conventions of Wilcox (1954), phe- rounded by clear, less calcic, oscillatory zoned rims. Glass inclu-
nocrysts are defined as greater than 0.3 m m in m a x i m u m dimension sions are least a b u n d a n t in the plagioclase of the most silicic sam-
and microphenocrysts as greater than 0.03 mm. ples, including the 1975-1976 andesite, Col 9. Following glass, apa-
Point counted modal analyses for each sample are listed in tite is the next most a b u n d a n t included phase. On average, apatite
Table 6 in order of increasing sample age. With the exception accounts for less than 0.05 vol.% of the plagioclase. It occurs
of hornblende, all phases present as phenocrysts and micropheno- as needles up to 0.1 m m in length, which are occassionally concen-
crysts are present in the groundmass (less than 0.03 mm). Both trated in crystal cores, but more commonly in distinct zones paral-
crystals and glass are included in the groundmass component of leling crystal margins. In the latter occurrence, the needles are
Table6. Owing to the microscoriaceous nature of Col 15, its usually oriented tangentially to the plagioclase rims. Any other
groundmass percentage is less certain. Significant features of the coexisting crystalline phase m a y be included in the plagioclase
data include (1) the mutually exclusive occurrence of stable horn- phenocryst rims, but these are rare.
blende and olivine, (2) the consistently high crystal contents (greater As is characteristic of calc-alkaline andesites, the plagioclase
than 0.03 m m diam.) in the suite, ranging up to 52 vol.%, and in these samples displays complex oscillatory zoning, which is
(3) the reduced hornblende contents (0.6 1.0 vol.%) and higher usually most pronounced on crystal margins, where more than
total crystal contents (47.2 51.6 vol.%) in the fourth cycle horn- 40 distinct zones m a y be present. C o m m o n l y a clear, oscillatory
blende-andesites, Col 9, 6, and 17, versus earlier hornblende-ande- zoned margin with a trend of decreasing calcium toward the rim
sites of similar bulk composition, Col 18, 2, and 7 (1.7-5.0 vol.% surrounds a more calcic core which is either (l) inclusion free,
hbd; 35.8 38.7 vol.% crystals). (2) separated from the clear margin by a penultimate zone of
glass inclusions, or (3) heavily riddled with brown glass inclusions.
Table 7 lists average microprobe analyses of plagioclase pheno-
cryst rims and the most sodic and most calcic point analyses in
Plagioclase each sample. Analyses are listed in order of increasing sample
age. All samples contain normally zoned plagioclase on average.
Plagioclase is the most a b u n d a n t crystalline phase in each sample. Microphenocryst analyses (unreported) of plagioclase and all other
The largest grains are generally greater than 3 m m in length, al- phases overlap with those of phenocryst rims in each sample. Neu-
though close inspection often reveals these to be annealed compo- tron activation analyses of hand picked plagioclase separates from
sites of several smaller grains. Rounded to irregular brown glass seven of the samples are given in Table 8.
inclusions are often present within the plagioclase and can be highly
concentrated in crystal cores or in one or more distinct zones
paralleling crystal margins. It is c o m m o n to find plagioclase crystals Orthopyroxene
with widely varying glass inclusion densities and styles within a
single thin section. A crude estimate of the volume percentage Orthopyroxene is usually the most a b u n d a n t mafic mineral in Co-
of glass inclusions within plagioclase is listed for each sample in lima's post-caldera andesites. It is generally less than 2 m m in
Table 7. Microprobe analyses of plagioclase
Col 9r 9ab 9an 6r 6ab 6an 17r 17ab 17an 15r 15ab 15an 18r 18ab 18an
SiO 2 55.95 57.02 51.57 55.25 56.72 53.43 54.58 59.36 45.86 54.65 57.27 51.83 54.02 57.93 51.50
AleO3 28.14 26.86 31.26 27.87 27.42 29.56 28.68 25.03 34.49 28.64 27.08 30.15 29.18 26.99 30.44
Fe203 0.53 0.53 0.54 0.57 0.53 0.59 0.55 0.59 0.41 0.55 0.51 0.52 0.57 0.47 0.51
BaO 0.03 0.05 0.00 0.03 0.06 0.00 0.04 0.12 0.01 0.03 0.06 0.00 0.03 0.05 0.01
SrO 0.09 0.02 0.04 0.08 0.03 0.07 0.11 0.05 0.11 0.09 0.13 0.07 0.10 0.00 0.13
CaO 10.11 8.82 13.02 9.90 9.42 12.14 10.51 6.53 16.87 10.92 9.52 13.90 11.64 8.90 13.03
NazO 5.82 6.62 4.31 6.13 6.29 4.87 5.65 7.95 2.20 5.28 6.00 3.75 5.18 6.62 4.43
K20 0.22 0.25 0.12 0.23 0.31 0.15 0.21 0.50 0.04 0.18 0.22 0.07 0.16 0.27 0.12
Total i00.89 100.17 100.86 100.06 100.78 100.81 100.33 100.13 99.99 100.34 100.79 100.29 100.88 101.23 100.17
Xan 0.484 0.418 0.621 0.466 0.445 0.575 0.501 0.304 0.807 0.528 0.461 0.669 0.549 0.420 0.615
Xab 0.504 0.568 0.372 0.522 0.538 0.417 0.487 0.669 0.t91 0.462 0.526 0.327 0.442 0.565 0.378
Xor 0.013 0.014 0.007 0.013 0.017 0.009 0.012 0.028 0.022 0.010 0.013 0.004 0.009 0.015 0.007
GI% 5 10 10 25 10
Col. 2r 2ab 2an 7r 7ab 7an 11 r 1 tab 11 an 13r 13 ab 13an
SiO2 54.57 55.66 45.39 54.28 56.33 51.94 52.88 54.73 49.11 53.58 56.57 47.52
AlzO3 29.47 28.02 35.43 29.14 27.07 30.85 29.64 28.38 32.34 29.52 27.43 33.86
Fe20 3 0.56 0.68 0.48 0.57 0.53 0.52 0.73 0.83 0.57 0.72 0.78 0.43
BaO 0.03 0.04 0.02 0.04 0.11 0.06 0.02 0.01 0.01 0.02 0.10 0.00
SrO 0.10 0.07 0.06 0.11 0.08 0.13 0.08 0.08 0.06 0.08 0.06 0.10
CaO 11.36 9.78 17.80 ll.01 9.02 12.62 12.12 10.36 15.14 12.11 9.43 16.75
Na20 5.26 6.02 1.75 5.50 6.74 4.50 4.93 5.68 3.30 4.89 5.65 2.41
K20 0.20 0.28 0.03 0.25 0.25 0.21 0.18 0.24 0.09 0.t8 0.74 0.04
Total 101.55 100.55 100.96 100.90 100.13 100.83 100.58 100.31 100.62 101.10 100.76 I01.11
Xa, 0.538 0.466 0.848 0.518 0.419 0.601 0.570 0.495 0.714 0.572 0.459 0.792
X,b 0.451 0.519 0.151 0.468 0.567 0.388 0.420 0.491 0.281 0.418 0.498 0.206
Xor 0.011 0.016 0.002 0.014 0.014 0.012 0.010 0.014 0.005 0.010 0.043 0.002
GI% 15 5 45 45
r: average of phenocryst rims; ab: most sodic point analysis; an: most calcic point analysis
G I % : Estimate of glass inclusion density in plagioclase
m a x i m u m dimension and invariably displays pink-green pleochro- sium enriched rims are most pronounced in the oldest of the sam-
ism. Titanomagnetite is a c o m m o n included phase in the rims pies, the olivine-andesites Col 11 and 13. In addition, approxi-
of both orthopyroxene and clinopyroxene, occurring as grains up mately 5% of the orthopyroxenes in these samples are jacketed
to 0.2 m m in diameter. Tiny (0.1 m m long) apatite needles and by thin clinopyroxene rims instead of by the Mg-rich orthopyrox-
glass are less c o m m o n as inclusions, although individual crystals ene rims. This phenomenon of ctinopyroxene jacketing is not
can be literally riddled with apatite. Similar to their occurrence observed in the hornblende-andesites. N e u t r o n activation analyses
within plagioclase, apatite needles and glass inclusions in pyroxenes of hand picked orthopyroxene separates from six of the samples
are often concentrated in crystal cores or in thin zones paralleling are listed in Table 10. The somewhat high N a contents for Col 6
crystal margins. and 17 orthopyroxene in Table 10 apparently reflect minor ground-
Table 9 lists average microprobe analyses of orthopyroxene mass contamination.
phenocryst rims and cores in order of increasing sample age. Total
compositional variation of orthopyroxene in a given sample is
never more than 10 mol% MgSiO3. Differences between average Clinopyroxene (Augite)
rim and core compositions are less than 2 tool% (Mg) for the
hornblende-andesites and less than 6 mol% (Mg) for the olivine- Augite is generally subordinate in a m o u n t to orthopyroxene, but
andesites. Only those samples belonging to the fourth eruptive usually predominates within crystal clusters as noted by Nelson
cycle (Col 9, 6, and 17) dominantly contain normally zoned ortho- (1979) at Volc~in Ceboruco. Augite is less than 2 m m in size as
pyroxenes, with relatively iron rich rims. Almost all orthopyroxenes phenocrysts, though slightly larger within clusters. Inclusion pat-
in pre-1961 samples have relatively m a g n e s i u m rich rims. Magne- terns are similar to those described for orthopyroxenes.
Table 8. Neutron activation analyses of plagioclase" P'~CN~ t ~ ' P -- ~D

ddd dd dd
Col 9 6 17 15 2 7 11
Na% 4.04 4.00 4.00 3.78 3.92 4.08 3.53

Sc 0.24 0.16 0.13 0.13 0.t7 0.16 0.28
Ti% <0.03 <0.03 <0.04 <0.06 <0.05 <0.05 <0.04
V <25 <30 <25 <18 <25 46 <22 d d d d d d d
Cr 0.5 < 1 <0.3 0.4 0.2 0.4 0.7
Mn 45 39 38 40 45 41 38
Fe% 0.39 0.38 0.39 0.36 0.41 0.40 0.47
M d ~ d N ~ d d d d d d
Ni <2 2.4 <2 <l 2.6 6.1 3.0
Zn 10 16 10 ii 17 10 9
Cs <0.2 <0.3 <0.3 <0.3 <0.4 <0.3 0.04
Ba 250 60 120 211 58 73 52 d d d d d d d
La 2.7 4.1 3.2 2.3 2.9 3.2 2.0
Ce 5.5 6.0 5.5 4.7 5.7 6.3 4.6
Nd 2.7 3.6 1.9 1.6 3.2 2.8 3.0 M d ~ N d ~ M d
Sm 0.34 0.46 0.33 0.27 0.32 0.38 0.31
Eu 0.72 0.71 0.85 0.71 0.64 0.70 0.49
Tb 0.04 0.03 0,03 0.03 0.03 0.04 0.03
Dy 0.2 0.2 0.3 <0.12 0.3 0.2 0.2
Yb 0.06 0.03 0.03 0.04 0.05 0.03 0.03
Lu 0.03 <0.03 <0.02 <0.02 <0.03 <0.03 <0.02
Hf 0.15 0.06 0.06 0.03 0.11 0.08 0.07
Th 0.02 <0.05 0.04 0.01 <0.02 <0.02 <0.02
a In ppm, except Na, Ti, and Fe; uncertainties as listed in Table 2 ,-A
~ 0 0 O 0 ~ ~
ddd dd
d
o
Table 10. Neutron activation analyses of orthopyroxene a ddd dd
Col 9 6 17 2 7 11
Na% 0.21 0.83 0.48 0.19 0.12 0.I3

Sc 37.9 34.7 36.2 32.6 35.4 35.8
Ti% 0.13 0.19 0.14 0.12 0.15 0.12 g
V 132 136 115 119 108 140
Cr 234 84 181 429 443 721
Mn 4,124 3,747 3,824 3,438 3,601 2,768 O
Fe% 12.9 11.2 11.9 12.0 12.4 11.2
sz~
Ni 140 79 131 254 237 400
Zn 266 248 257 239 244 188
Cs <0.6 <0.5 <0.6 <0.4 <0.4 0.5
Ba 23 172 98 38 31 28 8
La 1.9 5.9 3.9 1.7 0.9 <0.5
Ce 5.7 14.2 8.4 2.4 2.4 <0.8 U
Nd 3.5 9.5 5.4 2.2 2.4 1.2
Sm 0.95 1.96 1.27 0.41 0.42 0.27
Eu 0.22 0.47 0.30 0.11 0.11 0.09
Tb 0.20 0.29 0.26 0.13 0.13 0.09 O
Dy 1.8 2.4 1.6 1.2 t .0 0.9
Yb 1.14 1.64 1.33 0.81 0.93 0.72
Lu 0.20 0.28 0.23 0.15 0.16 0.11 ;>
Hf 0.3 1.4 0.7 0.2 <0.1 0.2 <
Th 0.21 0.84 0.30 0.12 0.16 0.13
I
.~C~ ooq
" In ppm, except for Na, Ti, and Fe; uncertainties as in Table 2 • 2 1 5 2 1.~o
5 ~
Table 11. Average microprobe analyses of clinopyroxene
Col 9r 9c 6 17r 17c 15r 15c 18 2 7 11r 11c 13r 13c
SiO 2 51.82 51.75 51.07 51.50 52.50 51.05 50.63 50.95 50.97 51.70 51.22 50.47 52.13 51.95
TiO2 0.46 0.59 0.56 0.67 0.44 0.58 0.69 0.56 0.55 0.49 0.66 0.70 0.60 0.49
AlzO3 2.54 2.67 2.21 2.78 1.79 3.61 3.71 2.37 2.86 2.19 3.61 3.91 2.36 3.43
FeO 9.04 8.63 8.91 8.84 7.56 7.34 7.91 9.04 8.68 8.60 7.81 8.77 8.22 6.24
MnO 0.03 0.29 0.30 0.28 0.22 0.24 0.21 0.29 0.25 0.27 0.16 0.21 0.23 0.17
MgO 14.75 15.18 15.48 15.21 16.29 16.02 15.68 15.30 15.31 15.43 15.76 14.77 16.58 17.06
NiO 0.01 0.01 0.02 0.01 0.00 0.02 0.02 0.00 0.01 0.01 0.03 0.04 0.00 0.02
CaO 20.21 20.23 20.14 20.08 20.66 20.13 20.29 20.43 20.56 20.43 20.22 20.18 19.82 20.46
Na20 0.47 0.45 0.44 0.50 0.42 0.41 0.44 0.42 0.48 0.46 0.48 0.51 0.35 0.37
Total 99.33 99.80 99.13 99.87 99.88 99.40 99.58 99.36 99.67 99.58 99.95 99.56 100.29 100.19
Xen 0.429 0.439 0.443 0.440 0.460 0.463 0.452 0.436 0.438 0.442 0.455 0.432 0.468 0.484
Xfs 0.148 0.140 0.143 0.143 0.120 0.119 0.128 0.145 0.139 0.138 0.126 0.144 0.130 0.099
Xwo 0.423 0.421 0.414 0.417 0.420 0.418 0.420 0.419 0.423 0.420 0.419 0.424 0.402 0.417
high Xen 0.478 0.452 0.495 0.485 0.449 0.441 0.459 0.473 0.507
low Xen 0.424 0.439 0.443 0.451 0.443 0.419 0.441 0.437 0.474
r: average of phenocryst rims; c: average of phenocryst cores
Table 12. Neutron activation analyses of ctinopyroxene a patterns in the olivine-andesites. Nearly all of the augite in Col 11,
like the orthopyroxene, is reversely zoned with well developed,
Col 9 6 17 2 7 11 Mg-rich rims. Augite in Col 13, however, is largely normally zoned.
Such dissimilar phenocryst zoning patterns are surprising in sam-
Na% 0.42 0.42 0.46 0.42 0.39 0.39 ples of nearly identical bulk composition (Tables 1-3). Neutron
Sc 109.0 112.3 112.5 112.5 116.8 110.7 activation analyses of hand picked augite separates from six of
Ti% 0.29 0.31 0.21 0.28 0.32 0.27 the samples are listed in Table 12.
V 275 299 275 288 273 339
Cr 640 584 650 734 504 1,245
Mn 2,176 2,207 2,210 2,228 2,320 1,873
Fe% 6.3 6.4 6.3 6.7 6.4 5.9 Hornblende
Ni 60 46 90 72 63 153
Zn 690 569 718 439 656 686
Basaltic hornblende with strong, reddish-brown pleochroism occurs
Cs <0.3 <0.5 <0.3 <0.4 <0.4 0.7
as large crystals up to 4 mm i.d. In all but the quickly quenched
Ba 25 37 20 < 20 11 24
scoria sample, Col 15, the hornblende is surrounded by either a
La 3.7 2.8 2.7 3.8 2.5 2.2
reaction rim of opacite (fine-grained magnetite) up to 0.15 mm
Ce t0.0 10.4 10.8 12.7 12.8 7.8
thick, or a granular reaction corona of pyroxene, plagioclase, and
Nd 10 14 12 14 13 10
magnetite. Both styles of reaction occur within a single thin section,
Sm 4.25 4.35 4.28 4.57 4.61 3.44
although reaction to opacite is restricted to microphenocrysts and
Eu 1.16 1.21 1.19 1.27 1.19 0.99
the rims of phenocrysts, suggesting that it is the initial stage of
Tb 0.77 0.89 0.77 0.90 0.88 0.69
hornblende decomposition. In many instances, hornblende crystals
Dy 6.0 6.1 6.1 6.0 6.3 5.2
have totally reacted to a granular, multiphase pseudomorph. These
Yb 2.80 2.81 2.74 2.80 2.97 2.36
aggregates are easily recognized as hornblende pseudomorphs on
Lu 0.39 0.39 0.38 0.39 0.44 0.31
the basis of their fine-grained, granular texture and external mor-
Hf 1.4 1.5 1.4 1.7 1.4 1.2
phology. They are quite unlike the coarse grained, often clinopyrox-
Th - <0.5 - <0.2 <0.5 <0.1
ene-dominated crystal clusters which are also present in all post-
caldera andesites at Volc/m Colima. Stewart (1975) suggested such
In ppm, except for Na, Ti, and Fe; uncertainties as in Table 2
coarse grained crystal clusters in andesites as possible hornblende
decomposition products. Textural patterns in Colima's horn-
Average microprobe analyses of augite phenocryst rims and blende-andesites do not support this interpretation. Garcia and
cores are listed in Table 11 in order of increasing sample age. Jacobson (1979) also reject hornblende decomposition as the origin
Compositional zoning is less pronounced in Colima's augites than of crystal clots in a suite of Cascade andesites. Hornblende micro-
in the orthopyroxenes, and if no significant rim to core zoning phenocrysts rarely survive eruptive reaction in Colima's lavas. The
was detected in a particular sample, a single composition is given absence of hornblende reaction relationships in the quickly
in Table 11. Total variation in each sample is always less than quenched scoria of Col 15 supports the concept that this style
5 tool% MgSiO3. Average rims and cores always differ by less of hydroxy-hornblende decomposition occurs during eruption and
than 2 mol% (Mg). Augite phenocrysts are normally zoned to cooling under low pressure conditions (Kuno, 1950). Inclusions
unzoned in the samples belonging to the fourth eruptive cycle of titanomagnetite, pyroxene, and plagioclase are occasionally pres-
(Col 9, 6, and 17). Olucr samples have normally zoned, unzoned ent in Colima's hornblendes, but generally the hornblende occurs
and reversely zoned augites. Of paricular interest are the zoning as large, inclusion-free crystals.
Table 13. Average microprobe analyses of hornblende Table 14. N e u t r o n activation analyses of hornblende and olivine"
Col 6 17 t5 18 2 7 Col 6h 17h 15h 2h 7h 11 ol
SiO2 42.41 42.50 42.61 43.99 43.15 42.65 Na% 1.96 2.02 2.02 2.03 2.01 0.05
TiO2 2.48 2.67 2.68 2.36 2.97 2.33 Sc 92.6 93.0 91.6 93.3 90.6 4.8
A1203 12.08 12.82 11.75 10.92 11.47 11.83 Ti% 1.52 1.56 1.64 1.70 1.50 < 0.02
FeO 12.78 10.60 12.26 11.89 11.83 12.04 V 655 706 738 709 639 25
MnO 0.18 0.15 0.18 0.25 0.24 0.17 Cr 205 244 186 206 359 974
MgO 13.79 14.94 13.98 14.65 14.23 14.54 Mn 1,317 1,266 1,353 1,369 1,317 1,716
NiO 0.02 0.03 0.03 0.01 0.00 0.01 Fe% 9.0 8.9 9.2 9.4 9.0 11.2
CaO II.I1 11.28 11.14 11.25 11.35 11.32 Ni 139 157 115 146 138 1,563
Na20 2.66 2.82 2.56 2.50 2.67 2.65 Zn 318 412 526 537 542 107
K20 0.38 0.40 0.41 0.37 0.45 0.40 Cs <0.4 <0.3 0.5 <0.2 <0.3 <0.3
F 0.12 0.13 0.08 0.10 0.11 0.12 Ba 128 118 133 111 113 <10
C1 0.05 0.02 0.03 0.03 0.04 0.03 La 2.7 2.9 3.0 3.1 3.0 < 0.3
H20 a 2.00 2.00 2.00 2.00 2.00 2.00 Ce 14.0 13.2 13.2 13.0 12.2 <0.5
Nd 17.4 16.2 17.1 19 16.7 <0.5
Total 100.06 100.36 99.71 100.32 100.51 100.09 Sm 5.95 5.34 6.03 6.34 5.54 0.041
Eu 1.88 1.77 1.96 2.01 1.79 < 0.04
high M g O 14.24 15.50 14.63 14.85 14.55 15.29 Tb 1.16 0.96 0.95 1.13 1.01 <0.07
low M g O 12.97 14.27 13.55 14.55 13.81 14.19 Dy 7.6 7.6 8.1 8.2 7.2 <0.4
Yb 3.44 3.14 3.42 3.43 3.21 0.08
X~ubdmmX 105 6.2 7.1 8.3 14.9 7.7 8.7
Lu 0.44 0.40 0.47 0.49 0.45 < 0.03
Hf 2.2 2.4 2.3 2.1 2.2 <0.07
a All H2O assumed equal to 2.00 wt.% Th 0.7 < 0.5 0.59 0.45 0.45 < 0.05
bd __ A M 7 T 8 V 2
X~cumm--(Xvac)(XMg) (XsI) (XoH)
h: hornblende; ol : olivine
a In ppm, except for Na, Ti, and Fe; uncertainties as in Table 2
Table 15. A v e r a g e m i c r o p r o b e analyses of olivine
Col 9x 17x llr llc llx 13r 13c 13x
SiO2 37.94 39.00 38.14 38.40 39.25 38.53 39.15 40.60

TiO2 0.05 0.00 0.03 0.02 0.02 0.02 0.02 0.02
FeO 23.76 18.01 24.28 22.94 11.66 22.15 19.49 12.57
MnO 0.43 0.32 0.40 0.37 0.15 0.38 0.29 0.19
MgO 37.69 42.76 37.44 38.70 47.25 39.47 42.07 47.60
NiO 0.14 0.09 0.07 0.10 0.37 0.10 0.14 0.24
CaO 0.09 0.09 0.17 0.15 0.12 0.15 0.14 0.15
Total 100.10 100.27 100.52 100.68 98.83 100.80 101.30 101.37
Xfo 0.739 0.809 0.733 0.751 0.878 0.761 0.794 0.871

Xfa 0.261 0.191 0.267 0.250 0.122 0.240 0.206 0.129
high Xeo 0.742 0.808 0.877 0.869
low Xfo 0.737 0.793 0.697 0.717
afo 0.546 0.654 0.538 0.563 0.771 0.578 0.630 0.759
afa 0.068 0.037 0.071 0.062 0.015 0.057 0.042 0.017
afo and afa calculated after Carmichael et al. (1977)

r: average of phenocryst rims; c: average of phenocryst cores;
x: average of cores of reacting xenocrysts
N o reliable analysis of phenocrystic hornblende was obtained Olivine

for CoI 9, the small percentage of hornblende in this sample having
almost completely reacted on eruption. Consistent rim to core
zoning was not detected in the analyzed hornblende of any sample. In Col 11 and 13, olivine occurs as rounded to euhedral microphe-
Average microprobe analyses of hornblendes and their ranges in nocrysts and phenocrysts up to 3.5 m m long, and also as resorbed,
M g O content are listed in Table 13 in order of increasing sample reacted xenocrysts. Olivine is also present in all of the hornblende-
age. N e u t r o n activation analyses of hand picked hornblende sepa- andesites as rare, r e m n a n t xenocrysts rimmed by pyroxenes. Oliv-
rates from five samples are listed in Table 14. ines contain only occasional titanomagnetite and glass inclusions.
Table 16. Average microprobe analyses of F e - Ti oxides
Col 9m 9h 6m 17m 15m 18m 2m 7m llm ~ 11m b t3m ~ 13m b
SiO2 0.06 0.00 0.14 0.10 0.04 0.10 0.11 0.13 0.19 0.19 0.18 0.15
TiOz 9.82 39.22 11.04 12.22 6.95 7.92 7.63 8.86 10.70 9.60 12.61 10.44
AI20 3 2.31 0.00 2.41 1.32 3.79 2.79 2.45 2.25 2.98 3.84 1.42 2.58
VzO3 0.58 0.20 0.51 0.50 0.60 0.52 0.56 0.54 0.78 0.62 0.75 0.64
Cr20 3 0.11 0.05 0.09 0.10 0.31 0.10 0.12 0.08 1.36 4.15 1.27 3.64
FeO 81.19 55.73 79.22 80.02 80.21 81.06 82.73 80.34 75.93 74.12 76.55 74.20
MnO 0.40 0.30 0.44 0.51 0.37 0.35 0.39 0.36 0.34 0.30 0.48 0.37
MgO 1.90 2.48 1.86 1.60 3.02 2.39 1.80 1.74 3.59 3.89 2.52 3.96
CaO 0.03 0.00 0.03 0.03 0.00 0.03 0.01 0.01 0.06 0.01 0.03 0.01
Total 95.82 97.98 95.23 95.90 94.71 94.75 95.24 93.76 95.15 96.09 95.04 95.36
Recalculated analyses
FezO 3 48.34 27.99 45.07 44.44 52.05 51.08 52.08 48.66 44.39 43.47 41.85 43.45
FeO 37.70 30.55 38.67 40.04 33.38 35.09 35.87 36.56 35.98 35.01 38.90 35.10
Total 101.24 100.78 100.26 100.85 100.51 100.38 101.02 99.19 100.38 101.08 100.00 100.34
high TiO2 10.38 41.23 12.99 14.45 8.20 10.28 8.72 10.64 14.10 10.34 13.81 ll.01
low TiO2 9.45 36.74 9.92 10.88 5.61 6.51 7.02 7.76 7.72 8.50 10.89 9.21
amt 0.667 0.627 0.619 0.714 0.707 0.722 0.686 0.607 0.586 0.582 0.593
m: titanomagnetite; H : hematite-ilmenite
a Titanomagnetite in groundmass and included in pyroxenes
b Titanomagnetite included in olivine
Table 17. Neutron activation analysis of titanomagnetite a in a single sample. Differences between average phenocryst rims
and cores in Col 11 and 13 are less than 3 mol% (Mg). Phenocrysts
Col 6 17 2 7 11 are normally zoned in both samples. Cores of reacting xenocrysts
are considerably more magnesian than phenocrysts. The neutron
Na% 0.37 0.53 0.39 0.52 1.55 activation analysis of a hand picked olivine separate from Col 11
Sc 29 30 18 24 28 is listed in Table 14.
Ti% 5.0 4.6 3.5 1.8 2.7
V 3,600 3,350 2,920 3,400 2,630
Cr 660 604 1,020 530 2,700 Fe-- Ti Oxides
Mn 3,700 3,770 - - 3,320
Fe% 52.3 49.0 56.0 50.0 35.1 Titanomagnetite is c o m m o n in all samples (1) as isolated microphe-
Ni 147 190 157 125 260 nocrysts and phenocrysts up to 1.0 m m i.d., (2) as inclusions within
Zn 805 740 680 730 390 the rims of pyroxenes and occassionally within plagioclase, olivine,
Cs <1 <0.7 <0.5 <0.5 <0.5 and hornblende, (3) within hornblende reaction rims, and (4) as
Ba 100 130 85 140 250 a b u n d a n t groundmass microcrystals. Hematite-ilmenite solid solu-
La 7.3 7.0 3.7 6.0 7.0 tions are extremely rare, owing to the low TiO2 contents
Ce 15 16 4.5 11 16 (0.60 0.79 wt.%) of these andesites. In Col 9, a small cluster of
Nd 12 9 5 8 10 hematite-ilmenite crystals was analyzed. A single, but often variable
Sm 2.4 2.32 1.07 2.0 2.0 population of titanomagnetite is present in the hornblende-andesite
Eu 0.47 0.49 0.23 0.35 0.57 samples. Titanomagnetite compositions in these samples could not
Tb 0.40 0.38 0.17 0.28 0.30 be consistently related to textural groupings. In the olivine-ande-
Dy 1.9 2.3 1.5 1.9 2.0 sites, however, titanomagnetites included within olivine pheno-
Yb 0.84 0.93 0.52 0.84 1.16 crysts are enriched in A1, Cr, and Mg, and depleted in Ti, V,
Lu 0.13 0.15 0.11 0.13 0.16 Fe, Mn, and Ca compared with other titanomagnetites in these
Hf 2.0 2.0 1.5 1.8 2.9 samples.
Th - <0.2 - <0.5 < 1.0 Average microprobe analyses for titanomagnetites, including
two populations in the olivine-andesites, and for hematite-ilmenite
a In ppm, except for Na, Ti, and Fe; uncertainties as in Table 2 in Col 9 are listed in Table 16 in order of increasing sample age.
The analyses have been recalculated for ferric iron after the method
of Carmichael (1967). N e u t r o n activation analyses of hand picked
Average microprobe analyses of olivine phenocryst rims and titanomagnetite separates from five samples are listed in Table 17.
cores and xenocryst cores in Col 11 and 13, and of remnant, xeno- Titanomagnetite in Col 11 and Col 13 is very fine grained and
crystic olivines in Col 9 and 17 are listed in Table 15. Total varia- restricted to the groundmass. Consequently, the Col I1 separate
tion in the olivine compositions is less than 18 mol% Mg/SiO4 is impure, as attested to by its high Na content.
Table 18. Ranges of weight fraction distribution coefficients
plag opx cpx hbd oliv titanmt
Sc 0.01 - 0.03 2.2 4.7 6.9 16 8.1 13 0.30 1.7 - 3.3

V 0.07 <0.47 0.47- 2.2 1.1 - 5.2 6.3 -13 0.08 8.7 54
Cr 0.02 <0.25 21 143 43 -245 31 90 34 93 -340
Ni <0.06 >0.61 >0.79 24 >4.6 9.0 7.2 16 58 9 . 6 - 19
Zn 0.13 0.25 2.6 4.4 6.5 12 5.0 8.7 1.5 5 . 4 - 13
Cs 0.04 <0.31 <0.31- 0.45 <0.17- 0.64 <0.15- 0.50 <0.27 <0.39-<0.63
Ba 0.08 - 0.36 0.03 0.23 0.02 0.05 0.15- 0.22 <0.02 0 . 1 ~ 0.40
La 0.13 0.21 <0.03 0.30 0.14 0.23 0.14 0.19 <0.02 0 . 2 ~ 0.45
Ce 0.12 - 0.18 <0.02- 0.33 0.20- 0.36 0.29- 0.37 <0.01 0.12- 0.42
Nd 0.08 - 0.16 0.05- 0.43 0.44- 0.70 0.70- 0.95 <0.02 0.25 0.55
Sm 0.07 0.10 0.06- 0.43 0.78- 1.3 1 . 2 - 1.7 0.01 0.29- 0.55
Eu 0.36- 0.73 0.07- 0.42 0 . 7 ~ 1.2 1.5 - 1.9 <0.03 0 . 2 ~ 0.42
Tb 0.05 0.08 0.15- 0.48 1.1 - 2.0 1.5 - 2.5 <0.11 0.37- 0.67
Dy <0.03 - 0.09 0.21- 0.56 1.2 - 1.9 1.7 - 2.5 <0.10 0.44 0.58
Yb 0.01 0.02 0.29- 0.56 0.93- 1.4 1.1 1.6 0.03 0.24- 0.47
Lu <0.05 <0.10 0.31 0.68 0 . 8 6 1.4 1 . 0 - 1.6 <0.08 0.32- 0.44
Hf 0.01 - 0.02 <0.02- 0.22 0.21- 0.34 0.34- 0.52 <0.02 0.30- 0.62
Th 0.004 <0.013 0.04- 0.22 <0.04-<0.16 <0.13- 0.25 <0.02 <0.05-<0.42
Distribution coefficients are calculated by dividing the concentration of an element in the crystal separate by that in the groundmass
separate. High and low values are reported for each element. Only a single olivine separate was analyzed
Distribution Coefficients bly precipitate in such subvolcanic magma chambers

(Anderson, 1976 ; Marsh, 1976). Phenocryst rims and
Weight fraction distribution coefficients (KD's) governing the parti- microphenocrysts in Colima's andesites are believed
tioning of trace elements between melt and crystalline phases can to have crystallized within the upper levels of the
be readily calculated from the neutron activation data in Tables 5,
8, 10, 12, 14, and 17. Ranges of Kv's for each element in each
volcanic system, just before or during eruption. The
analyzed phase of the suite are tabulated in Table 18. The listed equilibration pressure of the phenocryst assemblage
olivine and plagioclase KD'S compare favorably with those in Col 11 has been calculated as 1,000 bars (Ghiorso
compiled by Arth, (1976) for andesites and basalts. Pyroxene and and Carmichael, 1979), but the errors of the calcula-
hornblende KD's for rare earth elements generally range higher
tion are large. Consequently, calculations involving
in the Colima samples. The KD's for Ba are higher in Colima's
orthopyroxenes, but lower in the hornblendes than those of Arth's coexisting phenocryst rim compositions are
compilation. Because they account for the presence of important performed at assumed lithostatic pressures of 100,
included phases, such as apatite, in the analyzed minerals, distribu- 1,000, and 2,000 bars. Assuming an average density
tion coefficients determined from naturally occurring crystal- of 2.6 g/cm 3 for the volcanic pile, the pressure at
groundmass pairs may most realistically represent actual operative
magmatic values.
the base of the pile, 3 km below the summit is equal
to 765 bars.
Estimation of Intensive Variables

Temperature
Lithostatic Pressure
Table 20 lists pre-eruptive temperature estimates for
Of all the important intensive magmatic variables, Colima's post-caldera andesites derived from a variety
lithostatic pressure is one of the most difficult to of geothermometers. A brief discussion of each geo-
accurately estimate on the basis of phase equilibria. thermometer follows. Plagioclase/liquid geothermom-
No reliable geobarometer is currently available for etry will be discussed later in an evaluation of mag-
application to Colima's post-caldera andesites. There matic water contents.
are, however, geologic constraints on the pressures F e - Ti Oxide Geothermometry - Buddington and
of phase equilibria. The summit calderas at Volcfin Lindsley's (1964) F e - T i oxide geothermometer is
Colima and many other major composite volcanoes probably the most precise method presently available
undoubtedly record earlier voluminous eruptions for estimating magmatic temperatures and oxygen fu-
which significantly drained large magma chambers gacities (Wright and Weiblen, 1968 ; Helz, 1973 ; U1-
within or just below the volcanic piles. The majority mer et al., 1976). Titanomagnetite is present in all
of crystalline phases in calc-alkaline andesites proba- investigated post-caldera samples from Volc~in Co-
lima, b u t owing to the low TiOa contents o f these gated in this study. T e m p e r a t u r e s were c a l c u l a t e d
andesites (0.60-0.79 w t . % ) , h e m a t i t e - i l m e n i t e is ex- f r o m p h e n o c r y s t r i m c o m p o s i t i o n s , which p r e s u m a b l y
t r e m e l y rare. T h e only reliable analysis for this p h a s e r e c o r d m a g m a t i c t e m p e r a t u r e s on e r u p t i o n . T h e y
was o b t a i n e d for Col 9, the m o s t recent o f the sam- range f r o m 963~ ~ C in the h o r n b l e n d e - a n d e s i t e s
ples. F o l l o w i n g the r e c a l c u l a t i o n p r o c e d u r e s discussed a n d f r o m 1,017~ ~ C in the olivine-andesites.
in C a r m i c h a e l (1967), an e q u i l i b r a t i o n t e m p e r a t u r e Oxygen Isotope Geothermometry - Hand picked
o f 986 ~ C a n d oxygen fugacity o f 10 9.32 were calcu- m i n e r a l a n d g r o u n d m a s s separates f r o m C o l 9, 15,
lated for C o l 9. Because o f the e x t r e m e scarcity o f a n d 11 were k i n d l y a n a l y z e d for O18/O 26 b y K. K y s e r
h e m a t i t e - i l m e n i t e , these values are c o n s i d e r e d s o m e - w o r k i n g with J. O ' N e i l at the U S G S in M e n l o P a r k ,
w h a t unreliable. California. T h e d a t a are t a b u l a t e d in T a b l e 19. A n -
Pyroxene Geothermometery- Wood and Banno d e r s o n et al. (1971) f o r m u l a t e d a m a g m a t i c geother-
(1973) f o r m u l a t e d a g e o t h e r m o m e t e r b a s e d on the m o m e t e r b a s e d on the p a r t i t i o n i n g o f oxygen i s o t o p e s
shape o f the o r t h o p y r o x e n e - c l i n o p y r o x e n e miscibility between plagioclase a n d magnetite. K y s e r et al. (1980)
g a p for natural, i r o n - b e a r i n g systems. The geother- c a l i b r a t e d f o u r o t h e r isotopic exchange g e o t h e r m o m e -
m o m e t e r was later r e c a l i b r a t e d b y Wells (1977) using ters b a s e d on the d i s t r i b u t i o n o f oxygen isotopes be-
a larger b o d y o f e x p e r i m e n t a l data. B o t h o f the geo- tween coexisting p h e n o c r y s t s a n d / o r glass. T e m p e r a -
t h e r m o m e t e r s were a p p l i e d to the s a m p l e s investi- tures derived f r o m each o f these g e o t h e r m o m e t e r s
are listed in T a b l e 20.
C r y s t a l - M e l t G e o t h e r m o m e t e r s - G h i o r s o and
Tablel9.0x8/0~6Ratiosin mineralandgroundmassseparates C a r m i c h a e l (1979) recently f o r m u l a t e d a crystal-li-
Col plag cpx opx mt hbd oliv grdm quid e q u i l i b r i u m s o l u t i o n m o d e l b a s e d on e x p e r i m e n -
tally e q u i l i b r a t e d p l a g i o c l a s e - m e l t a n d olivine-melt
9 6.4 5.4 6.1 3.3 - - 6.9 pairs. C o m p o s i t i o n s o f plagioclase r i m - g r o u n d m a s s
15 6.4 5.5 5.5 4.0 6.7 - 5.7 pairs for C o l i m a ' s h o r n b l e n d e - a n d e s i t e s were used to
11 6.8 5.7 6.1 - - 5.7 6.5 solve for T-wt.% H 2 0 values at 1 a n d 3 k b pressure
As discussed in the text, the groundmass of Col 15 appears to (Table 20). As discussed in G h i o r s o a n d C a r m i c h a e l
be out of chemical and isotopic equilibrium with the crystals (1979) the presence o f olivine in C o l 11 allows de-
Analytical uncertainties (+ / - 0.1 ~ t e r m i n a t i o n o f a u n i q u e T - P - w t . % H 2 0 p o i n t for
Table 20. Temperature estimations for Colima's post-caldera andesites (~ C)
Col Ta T1 T2 T3 T4 T5 T6 T7 T8 T9
9 940 986 992 968 882 - - 880 1,020 1,021 (2.2)

986 (3.4)
6 940 - 995 974 . . . . . 987 (3.2)
970 (4.3)
17 940 - 995 977 . . . . . 952 (4.6)
15 1,000 - 1,023 1,010 1,028 - - 1,200 1,200 1,192 (0.2)

1,175 (0.4)
18 940 -- 976 971 . . . . . .
2 940 - 971 963 . . . . . 1,052 (2.5)

1,022 (3.8)
7 940 - 984 971 . . . . . 1,077 (1.9)
1,047 (3.2)
11 1,030 - 1,017 1,018 - 1,130 1,150 1,030 1,140 1,120 ( 0 . 8 )
13 1,060 - 1,056 1,053 . . . . . .
Ta:Assumed pre-eruptive temperature T6: O18/O16 cpx-oliv (Kyser et al., 1980)

TI: F e - T i oxides (Buddington and Lindsley, 1964) T7: O18/O16 cpx-gl (Kyser et al., 1980)
T2: Pyroxene phenocryst rims (Wells, 1977) T8: O18/O16 opx-gl (Kyser et al., 1980)
T3: Pyroxene phenocryst rims (Wood and Banno, 1973) T9: Intersections of albite and anorthite components at 1 and
T4: Ols/O 16 plag-mt (Anderson et al., 1971) 3 kb in T-%HzO space (Ghiorso and Carmichael, 1979)
T5: Ox8/O16 gl-oliv (Kyser et al., 1980) Numbers in parentheses are wt.% H20
equilibration of phenocryst rims of 1,120 ~ C, 1 kb, Oxygen Fugacity

and 0.8% H 2 0 .
F o r the sake of subsequent thermodynamic calcu- A single estimate of oxygen fugacity equal to 10 9.31
lations in which temperature is a variable, it is desir- at 986~ was obtained from the coexisting F e - T i
able to appraise the relative merits of the temperature oxide phases of Col 9 using the geothermometer of
estimates in Table 20 and to choose a single value Buddington and Lindsley (1964). In the olivine ande-
for each sample. In the case of the hornblende-andesites (Col 11 and 13), orthopyroxene, olivine, and titano-
sites, this process is greatly facilitated by knowledge magnetite were all stable pre-eruptive phases. Equa-
of the experimentally determined stability field of tion (1) can, therefore, be solved for pre-eruptive
hornblende. fo2 at assumed lithostatic pressures of equilibration.
The stability of hornblende in andesites and ba-
salts has been investigated experimentally by a variety 6MgzSiO4+ 6FeSiO3+ 02 = 2F%O4 + 12MgSiO3 (1)
of workers (Eggler, 1972a and b; Eggler and Burn- oliv opx gas titanmt opx
ham, 1973 ; Allen et al., 1975; Stern et al., 1975; Allen
and Boettcher, 1978). The upper temperature stability Thermodynamic calculations in this paper were
limit of hornblende decreases with increasing silica performed by fitting log K of each reaction be-
content of the liquid from 1,050 ~ C in basalts with tween 900 ~ and 1,100~ to an equation of the
5 0 w t . % SiO2 to 940 ~ 970~ in andesites with form log K = A / T + B + C ( P - 1 ) / T using the thermo-
5 9 w t . % SiO2. Accordingly, pre-eruptive tempera- dynamic data of Helgeson et al. (1978) and
tures in Colima's post-caldera hornblende andesites Ghiorso and Carmichael (1979). Coefficients for each
with 6 0 % - 6 2 % SiO2 (Col 9, 6, 17, 18, 2, and 7) are reaction are listed in Table 21. Equation 1 was solved
assumed to be 940 ~ C. Estimates for these samples at temperautes listed in Table 20 and pressures of
listed in Table 20 range from 880~ ~ C. For 100 and 1,000 bars, using activities of MgSiO3,
Col 15 (57.6% SiO2), temperatures calculated from FeSiO3, and Mg2SiO4 calculated from phenocryst rim
solid equilibria, both elemental and isotopic, range compositions listed in Tables 9 and 15. and activities
from 1,010~ ~ C. Liquid-solid geothermometer of Fe304 calculated from titanomagnetite composi-
of both types predict anomalously high temperatures tions listed in Table 16. A Vsoaias for the reaction is
for the hornblende-bearing Col 15 of 1,175~ ~ C. small and calculated values of pre-eruptive fo2 are
The groundmass glass of the 1913 scoria Col 15 ap- only weakly dependent on assumed pressures of equil-
pears to be out of chemical and isotopic equilibrium ibration.
with the crystalline phases. A pre-eruptive tempera-
ture of 1,000 ~ C is assumed for Col 15. Estimates for Col T (~ P (bars) logfo 2
the olivine andesites Col 11 and 13 range from 1,017 ~
1,150~ C and 1,053~176 C respectively. Pre-erup- ll 1,303 100 - 8.32
tive temperatures of 1,030~ and 1,060~ are 1,000 - 8.30
assumed for these samples. The assumed values of 13 1,333 100 - 7.83
pre-eruptive temperature are also listed in Table 20. 1,000 - 7.81
The choices are admittedly somewhat arbitrary, but
we are confident in the relative trend of the assumed Pre-eruptive T - f o 2 points calculated for Col 9 from
temperatures. F e - T i oxide geothermometry and for Col 11 and
Table 21. CoefficientsA, B, and C from the equation log K = A/T+ B + C(P- 1)/Tfor the listed reactions
Equation Reaction A B C
1 6MgzSiO4 + 6FeSiO3 + 02 = 2Fe304 + 12MgSiO3 27,081 -8.874 -0.20

oliv opx gas titanmt opx
FeSiO3 + 1/602 - 1/3Fe30~ + SiO2 3,916 -1.130 -0.114
opx gas titanmt liquid
1/3 Fe304 + MgTSisO22(OH)2= 7 MgSiO3 + FeSiO3 + 1/602 + HzO -8,603 8.650 0.142
titanmt hbd opx opx gas gas
Coefficients A and B were determined by fitting log K of each reaction at i bar pressure and temperatures between 900~ and 1,100~ C
Thermod~cnamic data is taken from Helgeson et al. (1978) except for SiO2 liquid which is taken from Ghiorso and Carmichael, (1979)
I I I I I
-8
-9
-I0
log fo2
-11
Fig. 7. T'-fo2 plot including data points for

-12 post-caldera andesites from V. Colima
J ~ Collland13 eqn.1 derived from Fe-Ti oxide compositions
and solution of Eq. (1), and data points for
-13 / 9 P~~ItdCsa"
~ide~r;roPmY
r~ e~eebo. . . .
post-caldera pyroxene andesites from V.
/ Nelson [19791 Ceboruco (Neson, 1979) derived from
I I I I I F e - 77 oxide compositions. Hbd-biot and
900 1000 1100 opx curves were taken from Carmichael
(1967). Eq. (2) is a best fit line through the
T% upper three data points
13 from solution of Eq. (1) are plotted on Fig. 7. The andesites are more alkalic, it must be concluded that
three points lie 1.2-1.6 log units above the N N O buf- fo2 was significantly higher in Colima's magmas (Luhr
fer, near the extrapolation of the curve for siliceous and Nelson, 1979). In this paper it is assumed that
magmas containing biotite and amphibole pheno- Colima's magmas are buffered in T - f o 2 space along
crysts in the absence of orthopyroxene (Carmichael the line through the three data points in Fig. 7, where-
and Ghiorso, 1980). by:
Fudali (1965) experimentally determined oxygen
fugacities in an olivine basalt, an olivine andesite, logfo2=-35.66+0.0209 (T ~ (2)
and two pyroxene andesites from the Mr. H o o d area,
by reproducing naturally occurring Fe + 3/Fe + 2 ratios This relationship is used in all subsequent calculations
at 1,200 ~ C. Determined values of fo2 ranged from on Colima's magmas. It is not greatly dependent on the
10 s.5 _ 10- 7, straddling the N N O buffer by approxi- assumed temperatures for the olivine andesites.
mately 0.7 log units. A nearly identical T - f o 2 rela-
tionship ( l o g f o ~ = N N O + / - 0.7) was calculated by
Nelson (1979) for andesites from Volc~in Ceboruco Silica Activity
using Buddington and Lindsley's (1964) F e - T i oxide
geothermometer. T - f o 2 points for post-caldera py- Each of the investigated andesites contains coexisting
roxene andesites from V. Ceboruco are also included orthopyroxene and titanomagnetite crystals, and if
in Fig. 7. The N N O buffer is essentially coincident a lithostatic pressure of equilibration is assumed, Eq.
with the curve for siliceous magmas containing ortho- (3) can be solved for the activity of silica in the coex-
pyroxene phenocrysts (Carmichael and Ghiorso, isting melt phase.
1980).
The relative positions of T - f o 2 trends for horn- FeSiO3+ 1/602 = 1/3Fe30r + SiOz. (3)
blende-andesites from V. Colima and pyroxene-ande- opx gas titanmt melt
sites from V. Ceboruco on Fig. 7 are supported by
a comparison of whole rock F e z O 3 / F e O + F e 2 0 3 ra- Reaction coefficients for Eq. (3) are listed in Table 21.
tios. Values for Colima's post-caldera andesites range The calculations were performed using FeSiO3 and
from 0.33-0.42 and are consistently higher than those Fe304 activities for orthopyroxene rim and titano-
from Ceboruco's andesites (0.21-0.25; Nelson, 1979). magnetite compositions (Tables 9 and 16), oxygen fu-
As discussed by Carmichael and Nicholls (1967), gacities as calculated from Eq. (2), and assumed pres-
Fe+3/Fe +2 ratios in silicate melts are direct functions sures for pre-eruptive equilibration of 100 and
of alkali content, fo2, and 1/T. Available geothermom- 1,000 bars. The results of the calculations are listed
eters record similar pre-eruptive temperatures for an- in Table 22. At 100 bars pressure, calculated values
desites from the two volcanoes. Because Ceboruco's of asio2meltrange from 0.65 0.74. As indicated, the calcu-
J.F. Luhr and I.S.E. Carmichaei: The Colima Volcanic Complex, Mexico. I 361
Table 22. Calculated silica activities Table 23. Calculation of PH2O from plagioclase geothermometry
Eq. 4 Ghiorso and Carmichael (1980) Col 9 6 17 15 2 7 11
Co1 T P
melt
aslo2
T P melt
asiQ T( ~ C) 940 940 940 1,000 940 940 1,030
(~ K) (bars) (~ K) (bars) PH~o (b.) 1,750 1,750 1,800 2,100 2,400 2,150 1,900
9 1,213 100 0.71 1,294 1,000 0.72 Water pressures calculated from the plagioclase geothermometers
1,000 0.59 of Kudo and Weill (1970) and Drake (1976) using assumed pre-
6 1,213 100 0.71 1,260 1,000 0.74 eruptive temperatures
1,000 0.59
17 1,213 100 0.69 1,225 1,000 0.75
1,000 0.58
plagioclase geothermometers can be used to predict
15 1,273 100 0.71 1,465 1,000 0.61
1,000 0.65
PH20 (Stormer and Carmichael, 1970). Groundmass
separates from seven of Colima's post-caldera ande-
18 1,213 100 0.65 - -- -
1,000 0.59
sites have been wet chemically analyzed (Table 4).
An estimate for pre-eruptive water pressure in these
2 1,213 100 0.65 1,325 1,000 0.7i
1,000 0.54
magmas was obtained by determining the value of
PH2o which is necessary to bring the temperature cal-
7 1,213 100 0.68 1,350 1,000 0.71
1,000 0.56
culated from plagioclase phenocryst rim-groundmass
pairs into harmony with the assumed value for pre-
11 1,303 I00 0.72 1,393 1,000 0.60
1,000 0.68
eruptive temperature. Calculated values of PH2o,
which range from 1,750 2,400 bars, are listed in Ta-
13 1,333 100 0.74 - - -
1,000 0.69
ble 23. The lowest values of PH2o are recorded for
the three samples of the most recent eruptive cycle
(Col 9, 6, and 17).
lation is strongly dependent on assumed equilibration Analyses of coexisting hornblende, orthopyrox-
pressure. At any value of P, however, ~melt ctSi02 is higher ene, and titanomagnetite in six of Colima's post-
for the three andesites of the most recent eruptive caldera andesites (no hornblende analysis was ob-
cycle (Col 9, 6, and 17), versus earlier hornblende- tained for Col 9) allow solution of Eq. (4) for water
andesites of similar bulk composition (Col 18, 2, and fugacity at assumed values for lithostatic pressure of
7). equilibration. Reaction coefficients for Eq. (4) are
Values ,~1
,,r "sio2
omelt were also calculated from analyzed listed in Table 21.
groundmass compositions at P = 1 kb as discussed in
Ghiorso and Carmichael (1979). These values and 1/3 Fe30 4 + MgTSisO22(OH)2
the calculated temperatures are listed in Table 22. Sil- titanmt hbd
ica activities calculated by this method are also higher = 7 MgSiO a + FeSiO a + 1/6 0 2 + H20. (4)
for Col 9, 6, and 17 than for the earlier, composition-
opx opx gas gas
ally similar Col 7 and 2. The significance of this obser-
vation is discussed later in this section in relation
to other peculiarities of the fourth cycle andesites. Since no thermodynamic data exists for monoclinic
MgTSisO22(OH)2 (cummingtonite), data for the or-
thorhombic variety (anthophyllite) have been used.
Water Fugacity and Water Content The difference in heat of formation of the two forms
should only be a few hundred calories. The greatest
Two methods of calculating magmatic water pressure uncertainty in the calculation of f ~ p by this method
can be applied to Colima's post-caldera andesites. lies in the determination of the activity of
The first is based on the empirical plagioclase geother- MgTSisO22(OH)2 in the hornblende. Powell (1975)
mometers of Kudo and Weill (1970) and Drake formulated a regular solution model for activity-com-
(1976), and the second is based on thermodynamic position relations in hornblendes based on coexisting
calculation. cummingonite-hornblende pairs in a suite of New
Plagioclase geothermometers are dependent on Zealand rhyolites, which record F e - T i oxide quench
coexisting plagioclase and silicate melt compositions, temperatures of 725~ ~ C. Using ahubmdm values cal-
and on the water pressure in the system. Given an culated from Powell's model, Eq. (4) predicts values
analysis of a coexisting plagioclase-melt pair and an of pre-eruptive water fugacity in Colima's andesites
independent estimate of equilibration temperature, that are considered unrealistically high, 3-4 kb. The
Table 24. Calculation of magmatic water content
Col T~ T~ P(bars) 7.hba

. . . = 31 7 ~bd
.... = 96
melt melt melt w t . oY0H20

melt w t . o~0H2
ma~ ma
f~o (b) Pn~o (b) XH20 f~n~o
elt (b) PH2Omelt (b) ff(HzOmelt w t . oyoH2Omelt w t . Ol/OH20magma
940 1,213 100 192 191 0.080 2.2 1.2 594 591 0.141 4.1 2.2
1,000 244 225 0.081 2.2 1.2 757 697 0.142 4.1 2.2
2,000 320 288 0.082 2.3 1.2 991 893 0.144 4.2 2.3
17 940 1,213 100 206 205 0.083 2.3 1.0 637 634 0.146 4.3 1.9
1,000 262 241 0.083 2.3 1.1 811 746 0.147 4.3 1.9
2,000 344 310 0.085 2.4 1.1 1,062 957 0.149 4.4 2.0
15 1,000 1,273 100 299 297 0.095 2.7 1.9 925 923 0.166 5.0 3.4
1,000 377 354 0.095 2.7 1.9 1,167 1,097 0.167 5.0 3.4
2,000 486 455 0.096 2.7 1.9 1,507 1,564 0.168 5.0 3.5
18 940 1,213 100 376 374 0.112 3.2 2.0 1,161 1,155 0.197 6.0 3.6
1,000 479 441 0.113 3.2 2.0 1,479 1,361 0.198 6.1 3.6
2,000 627 565 0.115 3.3 2.0 1,936 1,745 0.202 6.2 3.6
940 1,213 100 225 224 0.087 2.4 1.6 697 693 0.153 4.5 3.0
1,000 286 263 0.087 2.4 1.6 887 816 0.154 4.5 3.0
2,000 375 338 0.089 2.5 1.7 1,161 1,046 0.156 4.6 3.0
940 1,213 100 248 247 0.091 2.6 1.6 769 765 0.160 4.8 3.0
1,000 316 291 0.092 2.6 1.6 979 901 0.161 4.8 3.0
2,000 414 373 0.093 2.6 1.7 1,282 1,155 0.164 4.9 3.1
PI%o
melt is calculated from values o f f ~ and fugacity coefficients listed in Burnham et al. (1969)
m a j o r problem probably lies in trying to apply Pow- for the groundmass composition of the hornblende
ell's model outside of the compositional and tempera- andesites was used in the calculation. Pre-eruptive
ture ranges of its formulation. Allen and Boettcher water contents for the m a g m a s (whole rock) can then
(1978) recently published analyses of two coexisting be determined by weighting the water contents of
o r t h o p y r o x e n e - h o r n b l e n d e pairs f r o m experiments on the melt phase and hornblende crystals by the volume
a Mt. H o o d andesite at 13 kb, 900 ~ C, and x~IaPo= fractions of groundmass and hornblende in each sam-
0.75, and 13 kb, 940 ~ C, and X ~ = 0 . 5 0 , with fo~ ple (Table 6). Values of pre-eruptive f ~ P , pv, • p yIxmHe l2t0 ,
buffered along H M . A n unanalyzed o p a q u e mineral, wt.% U 2 0 in melt (groundmass), and wt.% H 2 0
u n d o u b t e d l y titanomagnetite, was reported in both in m a g m a (whole rock) calculated for Plith=100,
runs. Converting X~I~p to f~i~) using the modified Red- 1,000, and 2,000 bars, and 7 .hbd ... -- -- 9 6 and 31 are listed
lich-Kwong equation of Flowers (1979), formulating in Table 24.
yhbd
~cumm, ~opx
~en, and ,~fs"~ f r o m the reported analyses as Values offn2o calculated from Eq. (4) are strongly
discussed in Tables 13 and 9, and assuming amt=0.75 dependent on assumed lithostatic pressures. Conver-
in both runs, which has only a small effect on the sion to v~eltZXH20 (Carmichael et al., 1977) has a similar,
calculation, Eq. (4) was solved for the activity coeffi- opposing pressure dependency, however, resulting in
cient of cummingtonite in hornblende for the two wt.% H 2 0 values that are only weakly dependent
experimental runs. The calculated values are .cumm'~hbd= on assumed pressure. Estimates of pre-eruptive mag-
96 and 31, for the 900 ~ and 940 ~ runs respectively. matic water content in Colima's post-caldera horn-
These extracted activity coefficients have been mul- blende-andesites range f r o m 1.0-3.6 wt.% as listed
tiplied by values of X ~ m listed in Table 13 to derive in Table 24. At present, this m e t h o d of calculation
two estimates for a~ba~m in each sample. Equation is probably only suitable for estimates of relative,
(4) was then solved for f~ia~ using c o m p o n e n t activities n o t absolute m a g m a t i c water contents.
calculated f r o m o r t h o p y r o x e n e rim and titanomagne- Values of Pn2o determined f r o m plagioclase geo-
tite compositions at assumed lithostatic pressures of t h e r m o m e t r y (Table 23). are s o m e w h a t higher than
equilibration of 100, 1,000, and 2,000 bars. Values water pressures calculated f r o m Eq. (4) at P~ith=
of f ~ can be converted to mole fractions o f H2O 2,000 bars and y~b~mm= 96 (Table 24). This discrepancy
in the melt phase \-,x rvmemt~
H 2 0 / and ultimately to wt.% H 2 0
m a y be due to the fact that plagioclase-melt equilibria
in the melt as discussed by Carmichael et al. (1977). experiments were conducted at PH2O= PxotaV At un-
A n average gram formula weight of 69, determined dersaturated conditions, which p r o b a b l y apply in
J.F. Luhr and I.S.E. Carmichael:The ColimaVolcanicComplex,Mexico. I 363
most natural magmas, these two variables have op- activities of silica. The lack of slight Mg-rich rims
posing influences on plagioclase-melt equilibria. De- on orthopyroxenes in the fourth cycle andesites is
spite absolute differences in calculated water contents, also interpreted as consistent with relatively lower
both the plagioclase geothermometers and Eq. (4) water contents, as discussed in the next section.
predict the same relative trends, with the lowest, pre-
eruptive water pressures for Col 6 and 17, samples
from the most recent eruptive cycle. As indicated in
Table 24, the calculated magmatic water contents of Mineral Zoning and Textural Patterns
these samples are 0.4-1.7 wt.% lower than those of
earlier hornblende-andesites. The consistency of the A characteristic feature of calc-alkaline andesites is
two methods of calculation encourages us in the inter- their textural complexity, often accompanied by com-
pretation that magmatic water content has been signi- plex mineralogical zoning patterns. In Colima's post-
ficantly reduced during the most recent eruptive cycle. caldera andesites, the most striking mineral zoning
and textural features are the Mg-rich rims on pre-1961
pyroxenes and the brown glass inclusion zones within
plagioclase phenocrysts. The same phenomena have
Andesites of the Fourth Eruptive Cycle been reported in the literature from a wide spectrum
of calc-alkaline provinces (Kuno, 1936, 1950; Mac-
The third eruptive cycle at Volc/m Colima began in Donald and Katsura, 1965; Eichelberger, 1975, 1978;
1869 with the eruption of hornblende-rich block lavas. Anderson, 1976; Johnston, 1978; Sakuyama, 1979).
Col 2 and 18 were sampled from these flows and Using optical methods, Kuno (1936) identified rel-
record the highest pre-eruptive water contents in the atively Mg-rich rims on pyroxenes and clear, rela-
post-caldera suite (Tables 23 and 24). The third cycle tively Ca-rich rims on glass inclusion riddled plagio-
ended in 1913 with a voluminous ashflow eruption clase cores in the pigeonite-bearing andesites from
of hornblende-rich andesite (Col 15), also recording Hakone Volcano, Japan. The observation that all
a high water content. phenocrystic phases in these andesites posess higher
The three investigated andesites which erupted temperature rims than cores led Kuno to invoke a
since 1961, during the most recent eruptive cycle pre-eruptive mixing event involving a dacitic magma
(Col 9, 6, and 17), display a number of peculiarities and a hotter andesitic magma. Phenocryst cores were
relative to earlier hornblende-andesites of nearly thought to have crystallized in the cooler dacitic end
identical bulk composition (Col 18, 2, and 7). These member; the andesitic magma was presumably crys-
include: (1) lower hornblende contents (0.6-1.0% vs tal free. Most subsequent investigations cited above
1.7-5.0%), (2) higher total crystal contents have supported Kuno's early interpretations. Fol-
(47.2-51.6% vs 35.8-38.7%), (3) higher pre-eruptive lowing crystal growth rate studies by Lofgren et al.
silica activities as calculated from Eq. (3) (0.69-0.71 (1974, 1975) and Donaldson et al. (1975), however,
vs 0.65-0.68 ; assuming Plith= 100 bars), (4) lower pre- petrologists recognized similarities between glass in-
eruptive magmatic water contents as calculated from clusions incorporated by skeletal crystals grown under
Eq. (4) (1.9 2.2% vs 3.0-3.6%; assuming Plith= supercooled conditions and many glass inclusions in
100bars and ~)hba. . . . -_- 9 6 ) , and (5) relatively iron rich natural crystals (Gutmann, 1977; Dungan and
rims on orthopyoxenes. All of these observations are Rhodes, 1978; Evans and Nash, 1979; Rhodes et al.,
thought to be consistent with the interpretation that 1979). Dungan and Rhodes (1978) outlined criteria
andesites of the fourth eruptive cycle were signifi- for distinguishing resorption from primary growth
cantly less hydrous than earlier post-caldera andesites inclusions.
of similar bulk composition. The highly energetic ash-
flow eruption of 1913 (Waitz, 1915), which closed
the third eruptive cycle, may have triggered this tran-
sition by depleting Colima's magma system in vola- Olivine-Andesites
tiles. At constant bulk composition and temperature,
less hydrous magmas will have higher solidus temper- The most pronounced Mg-rich pyroxene rims and
atures, and thus, higher crystal contents. Silica activ- clear rims on glass inclusion riddled plagioclase cores
ity was higher in the fourth cycle andesites as the in Colima's post-caldera suite occur in the olivine
combined result of higher crystal contents leading andesites, Col 11 and 13. It might be suggested that
to more siliceous groundmasses (Table 4) and of lower both of these rim phenomena occurred on eruption
water contents. Water acts to decrease polymerization in a manner analogous to the hornblende reaction
in silicate melts and lower water contents cause higher rims. Both phenomena are observed in the quickly
364 J.F. Luhr and I.S.E. Carmichaeh The Colima Volcanic Complex, Mexico. I
Composite Phenocryst Textures and

Zoning Profiles in the Olivine Andesites
OLIV OPX CPX PLAG
Col 11
.;':tA
Xfo Xen _ _ Xen
.59
Xlln
Col 13
::tA .:':t/ :::tA Fig. 8. Idealized phenocryst textures and
zoning profiles in the major silicate phases
of the olivine-andesites Col 11 and 13
quenched scoria of Col 15, however, whose horn-

phen. rims phen. cores
blende does not show reaction rims, demonstrating
that the anomalous pyroxene and plagioclase rims oliv opx oliv opx
record pre-eruptive conditions. Idealized zoning pro-
files for the four major phases in Col 11 and 13 are Col 11 1.21 1.25 1.11 1.48
Col 13 1.05 1.15 0.87 1.58 .
shown in Fig. 8. As mentioned previously, olivine and
plagioclase phenocrysts in both samples and most
clinopyroxenes in Col 13 are normally zoned, while
most orthopyroxene phenocrysts of both samples and however, predict radically different liquid ratios
clinopyroxenes of Col 11 are reversely zoned, with in both samples. Mysen (1975) and Bender et al.
thin, Mg-rich rims. (1978) have criticized the conclusions of Roeder
In an experimental study of equilibria between and Emslie (1970), demonstrating slight temperature
olivine and basaltic liquids, Roeder and Emslie (1970) and pressure dependencies to the olivine/liquid
suggested that the distribution coefficient for parti- KD. No attempt has yet been made to experi-
tioning of ferrous iron and magnesium between oliv- mentally redetermine orthopyroxene/liquid KD rela-
ine and liquid, KD --- u 1-'J"
7FeO
6 Ix MgO"x
v oMgO
l i v v"xUFeO,
q is equal to tions, but it seems improbable that temperature and
0.30 and is essentially independent of temperature. pressure dependencies of the orthopyroxene KDwould
Combining their data with that of Medaris (1969) be in the opposite sense to those for olivine. As a
and Williams and Eugster (1969) on the partitioning result, it must be concluded that the cores of olivine
of Fe + 2-Mg between orthopyroxene and olivine, they and orthopyroxene phenocrysts in Col 11 and 13
also calculated a distribution coefficient for Fe + 2-Mg were never in equilibrium, and must have crystallized
partitioning between orthopyroxene and liquid of in two different magmatic environments.
0.23, similarly independent of temperature. These re- The most probable explanation for the origin of
lationships can be used to demonstrate the presence Colima's post-caldera olivine-andesites, therefore, is
or absence of equilibrium between coexisting orthopy- through the mixing of a basic, olivine-bearing magma
roxene and olivine crystals, and for complexly zoned with a less basic, orthopyroxene-bearing magma prior
crystals, they can be used to determine whether or to eruption. Titanomagnetites included within olivine
not these two phases have maintained equilibrium phenocrysts of these samples are compositionally dis-
throughout the crystallization interval (Johnston, tinct from those included within other phases or in
1978). Values of Xvco/XMgo in the liquid predicted the groundmass (Table 16), supporting an indepen-
by the above expressions for equilibrium with olivine dent origin for the olivines. Normally zoned clinopy-
and orthopyroxene phenocryst rim and core composi- roxenes, abundant in Col 13, also presumably crys-
tions in Col 11 and 13 are listed below. The liquid tallized in the more basic, end member magma, while
values predicted by olivine and orthopyroxene rim the reversely rimmed clinopyroxenes dominating in
compositions are nearly identical in both samples, Col 11 crystallized within the more siliceous, orthopy-
suggesting that the rims of these phases crystallized roxene-bearing magma. Temperatures of 978 ~ and
in equilibrium. The phenocryst core compositions, 975~ were calculated from the pyroxene geother-
mometers of Wells (1977) and Wood and Banno observed on all orthopyroxenes in pre-1961 horn-
(1973) respectively for the average core compositions blende-andesites. As discussed earlier, the post-1961,
of the reversely zoned pyroxenes of Col 11. These fourth cycle andesites were probably less hydrous
temperatures are thought to reflect the pre-mixing than previous andesites of similar bulk composition
temperature of the more siliceous end member which do show Mg-rich pyroxene rims (Col 18, 2,
magma. The values are similar to calculated pyroxene and 7). This observation suggests an alternative to
temperatures in Col 9, 6, 17, 18, 2, and 7, implying that magma mixing as an origin for Mg-rich pyroxene
the more siliceous end member was a typical, 60%- rims.
62% SiO2 andesite. It must first be assumed that some of the water
Textural and compositional relationships in the in Colima's andesites is inherited shortly before erup-
plagioclase of Colima's olivine andesites cannot be tion. This may occur as groundwater convects into
completely reconciled with the postulated magma a magma chamber immediately below the volcano,
mixing origin. As noted by Kuno (1950) for the pla- or as hydrothermally altered portions of the volcanic
gioclase at Hakone Volcano, the included glass in pile are assimilated by the magma. An observation
Colima's plagioclase is compositionally indistinguish- made by Kuno (1950) supports the assumption of
able from groundmass glass. It is, therefore, unlikely a late stage addition of water. In his classic study
to have formed from closed system heating and re- of Hakone Volcano and its vicinity, Kuno noted that
crystallization of the plagioclase. Kuno suggested Pliocene volcanics which erupted through a 4-km-
open system heating involving diffusion between the thick sequence of marine clastics in the northern part
plagioclase and melt. The plagioclase in Col 11 and of the region commonly contain hornblende pheno-
13 is compositionally different from most texturally crysts, whereas volcanics erupted to the south, away
similar plagioclase discussed in the cited literature from this marine basin, are almost devoid of horn-
in that the clear rims surrounding the glass inclusion blende. Kuno's observation implies that hornblende
riddled core areas are less calcic than the cores. The was stabilized in the northern magmas by a near sur-
core plagioclase has not been totally recrystallized face addition of water from the country rock. A simi-
to this calcic composition, because phenocrysts with lar near surface addition of water may occur at Vol-
only a penultimate zone of inclusions also have calcic can Colima, and may be responsible for the relatively
cores. The glass inclusion riddled core areas of Col 11 hydrous nature of its magma system. A detailed study
and 13 plagioclase probably reflect a period of skeletal of the oxygen isotopic ratios in andesites from Colima
crystal growth and melt incorporation under su- and other Mexican volcanoes is in progress and may
percooled conditions. Supercooling may have resulted ultimately constrain the origin of water in these
from the inferred magma mixing event, but presently magmas. As water is added to a magma, its dissocia-
unexplained aspects of the scenario include: (1) the tion will cause an increase in magmatic fo2. This will
location of the plagioclase crystals prior to mixing; oxidize a greater amount of the iron in the liquid
high or low temperature end member magma, (2) to Fe § In response, all phases will try to reach
the cause of the transition from glass inclusion riddled equilibrium with the lowered Fe+2/Mg ratio of the
cores to clear rims, and (3) the relative timing of liquid, and will do so by precipitating relatively Mg-
events recorded in plagioclase crystals and those rich rims. This effect will be most pronounced in
recorded in orthopyroxene crystals. The proposed olivine and orthopyroxene owing to their high MgO
mixing event presumably induced heterogeneous contents. In the case of the post-1961, fourth eruptive
nucleation of augite on orthopyroxene, resulting in cycle andesites, it is hypothesized that the near surface
the augite jackets observed in the olivine-andesites. addition of water was restricted, resulting in lower
Woodruff and Anderson (1979) observed similar au- magmatic water contents. Consequently, the pyrox-
gite jackets in olivine-andesite scoria from a satellite enes did not need to precipitate Mg-rich rims in order
vent of Pavlof Volcano, Alaska. The Pavlof scoria to remain in equilibrium with the melt. Reverse zon-
also contain ample evidence of pre-eruptive mixing. ing in pyroxenes may, then, reflect increasing or con-
stantfo2 during crystallization, and not an actual in-
crease in temperature with time.
Plagioclase crystals in Colima's hornblende ande-
Hornb lende-Andesites
sites commonly contain concentric bands of brown
glass inclusions. These also probably reflect periods
Zones of brown glass inclusions in plagioclase pheno- of rapid crystal growth, but the exact cause of this
crysts and Mg-rich pyroxene rims are also present, condition is unclear. In a single thin section, some
though less pronounced, in Colima's post-caldera crystals show numerous inclusion bands while others
hornblende-andesites. Slight Mg-rich rims are show none at all. It is considered significant that the
366 J.F. Luhr and I.S.E. Carmichael: The Cotima Volcanic Complex, Mexico. I
percentage of glass inclusions in plagioclase generally r*o*r~%

correlates with the SiO2 content of the whole rock
samples (Tables 1 and 7). Whatever the cause of the
glass inclusions, the effect becomes less obvious in
more evolved magmas.
Sequence of Crystallization
in the Hornblende-Andesites
px
In Colima's andesites, plagioclase phenocrysts con-
tain tiny apatite needles, but rarely inclusions of other
silicate or oxide phases. Consequently, plagioclase is
interpreted as the liquidus silicate phase in each
magma. The needle-like morphology of included apa-
tites within both plagioclase and pyroxene pheno-
crysts, and their common orientation parallel to phe- Na20+K20 MgO
nocryst margins suggest that apatite crystallized meta-
Fig. 9. AFM diagram showing analyzed whole rock and groundrnass
stably adjacent to growing crystal faces of other points connected by tie lines, ranges of phenocryst core composi-
phases and was not itself a stable phase in the tions, and trends for the tholeiitic Skaergaard (Wager, 1960) and
magmas. Titanomagnetite inclusions in Colima's py- Thingmuli (Carmichael, 1964) suites and the calc-alkaline suite
roxenes are largely restricted to crystal rims, implying from Mt. Jefferson (McBirney, 1969). wt.%
that the pyroxenes preceded titanomagnetite in crys-
tallization. In support of this observation, all avail-
able experimental data on andesitic systems (Eggler, sites generally have the largest grain size of any phase,
1972a; Eggler and Burnham, 1973) demonstrate that despite the fact that they are considered to have crys-
titanomagnetite is first stable at least 20~ below tallized through the smallest temperature interval
the upper temperature of orthopyroxene stability for prior to eruption. This observation demands a high
all realistic values of magmatic fo2 (log fo2 = NNO + / crystal growth rate for hornblende relative to the
-1.5). These experiments also show orthopyroxene other phases.
to precede clinopyroxene in order of crystallization.
Since the inclusion patterns of Colima's orthopyrox-
enes and clinopyroxenes are indistinguishable, it is Compositional Variations Within the Suite
concluded that both pyroxenes preceded titanomag-
netite. The abundance of titanomagnetite crystals in Whole rock and groundmass compositions of ana-
pyroxene phenocryst rims and their essential absence lyzed, post-caldera samples are plotted on an AFM
in plagioclase rims implies a nucleation preference, diagram in Fig. 9, with each pair joined by a tie line.
which may seriously complicate the conventional ap- The compositions of olivine, clinopyroxene, horn-
proach to determining sequences of crystallization blende, and titanomagnetite phenocryst cores in these
from inclusion patterns. Hornblende in Colima's an- samples are also indicated, as well as trends for the
desites occurs as large crystals which are commonly tholeiitic suites from Skaergaard (Wager, 1960) and
free of other phases, but do contain titanomagnetite, Thingmuli (Carmichael, 1964) and the calc-alkaline
pyroxene, and plagioclase inclusions. Experimental suite from Mt. Jefferson, a Cascade volcano (McBir-
determination of phase relations in Paricutin and Mt. ney, 1969). Whole rock-groundmass tie lines for
Hood andesites by Eggler (1972a) and Eggler and the hornblende-andesites are subparallel to one an-
Burnham (1973) demonstrate that hornblende is pre- other, to the locus of hornblende-andesite whole rock
ceded by plagioclase and pyroxenes for all H20 un- points, and to the whole rock trend for Mt. Jefferson.
dersaturated conditions below 5 8 kb confining pres- The whole rock-groundmass tie line for the olivine-
sure. Petrographic and experimental observation are andesite Col 11, has a distinctly different orientation,
consistent, then, with the interpretation that horn- which simply reflects the presence of olivine rather
blende precipitated last in the crystallization sequence than hornblende and titanomagnetite as a phenocrys-
of Colima's andesites, following titanomagnetite. tic and microphenocrystic phase.
Kuno (1950) reached a similar conclusion regarding As seen in Fig. 9, the order of decreasing Fe/Mg
the timing of hornblende crystallization in Hakone in Colima's phenocrystic phases is titanomagnetite
andesites. Hornblende phenocrysts in Colima's ande- hornblende - orthopyroxene - clinopyroxene - oliv-
J.F. Luhr and I.S.E. Carmichael: The Colima VolcanicComplex, Mexico. I 367
ine. Colima's post-caldera andesites display the lack nor participation by titanomgnetite (0.5 1.0 wt.%),
of iron enrichment which is characteristic of calc- do not encounter severe difficulties in reproducing
alkaline, orogenic suites. If magmas belonging to such vanadium contents in Colima's post-caldera suite.
a suite are genetically related to one another by crys- Greater difficulties are encountered with other ele-
tal-liquid equilibria, such as equilibrium partial melt- ments, particularly the highly compatible elements
ing or crystal fractionation, then a major role must Ni and Cr. Experimental arguments against an impor-
be played by a residual or fractionating mineral with tant role for titanomagnetite fractionation are based
a high Fe/Mg ratio. Candidates that have been on the observation that titanomagnetite is not a near-
suggested for this role include titanomagnetite (Os- liquidus phase in calc-alkaline andesites and basalts
born, 1962), hornblende, and garnet (Green and Ring- under realistic fo2 conditions (Eggler, 1972a; Eggler
wood, 1968). The data plotted on Fig. 9 qualitatively and Burnham, 1973). Petrographic observation of Co-
demonstrate that it is possible to generate Col 15 or lima's andesites supports a relatively late (post-pyrox-
17 from Col 11 through separation of observed oliv- ene) position for titanomagnetite in the crystallization
ines and pyroxenes, but that even the analyzed horn- sequence. The stability of spinels in calc-alkaline
blendes are not sufficiently iron-rich to generate the magmas may, however, be significantly enhanced at
trend of the hornblende-andesites by separation. Gar- higher pressures (Garcia and Jacobson, 1979).
net is not observed in any of Colima's andesites and Assuming that Colima's post-caldera suite is dif-
is, therefore, untestable in this regard. As noted by ferentiated by separation of observed phases, titanom-
Wright (1974) and Barker (1978), normalization agnetite appears to become an important, participat-
procedures employed in the construction of AFM ing phase by the time magma compositions evolve
diagrams distort compositional trends, such that crys- to that of Col 15 (57% SiO2). Significant in this re-
tal fractionation in Colima's magmas can not be gard is the essential absence of phenocrystic and mi-
quantitatively assessed from Fig. 9. Major element crophenocrystic titanomagnetite in the olivine ande-
least squares fractionation models (Wright and Do- sites (0-0.7 vol.%) and its presence in greater amounts
herty, 1970) discussed below allow the derivation of (0.8-2.5 vol.%) from Col 15 to more evolved compo-
Col 15 from Col 11 by separation of olivine and py- sitions (Table 6).
roxenes. Models involving only hornblende+plagio- Realistic geochemical models for the origin of
clase separation from Col 15 define a trend slightly calc-alkaline volcanic suites must be capable of simul-
displaced from the Col 15-17-9 trend towards the Fe taneously treating a variety of processes, including
apex. Hornblende fractionation is significantly more magma mixing and crystal fractionation (O'Hara,
effective than pyroxene fractionation in suppressing 1977). Such complex modeling of an active, dynamic
Fe-enrichment, but exact duplication of the whole volcanic system is considered beyond the scope of
rock major element trends from Col 15-9 requires mi- this paper. In this section, we test two very simple
nor coprecipitation of titanomagnetite. Hornblende end member models for the origin of compositional
controlled fractionation also seems unlikely owing to variations in Colima's post-caldera suite, crystal frac-
the late position of hornblende in the crystallization tionation and magma mixing.
sequence of Colima's andesites.
The early progressive enrichment of Fe relative
to Mg displayed by tholeiitic suites on AFM diagrams Crystal Fractionation Models
is halted during intermediate stages as titanomagne-
tite becomes a near liquidus phase driving the trend In order to rigorously test simple crystal fractionation
away from the Fe apex (Carmichael, 1964). Objec- of observed phenocrystic phases as the mechanism
tions to an important role for titanomagnetite in sup- producing the observed variations, a two-stage ap-
pressing early Fe-enrichment in calc-alkaline suites proach was employed (Nelson, 1979). In the first
have been made on geochemical and experimental stage, a major element, linear least squares computer
grounds. Taylor (1969) argued that separation of suf- program (Wright and Doherty, 1970) was used to
ficient titanomagnetite to inhibit iron enrichment calculated the proportions of phases which must be
would result in prohibitively low vanadium concen- removed from the chosen parent magma in order
trations in derivative magmas. Since Taylor's study, to derive the chosen daughter magma. P20~, which
a number of calc-alkaline suites have been shown to does not vary significantly through the suite, was not
record pronounced decreases in vanadium with in- included as a component in the calculations. Analyzed
creasing silica (Garcia and Jacobson, 1979). Titano- compositions of phenocryst cores in the presumed
magnetites in analyzed andesites from Volcfin Colima parent were used in each case. The quality of the
contain 2,600-3,600 ppm vanadium. Crystal fraction- fit is indicated by the sum of the squares of the
ation models presented in this section involving mi- observed minus calculated values for each component
Table 25. Three of the most successful crystal fractionation models assuming total equilibrium for the trace elements (Arth, 1976)
Col 13 to Col 15 Col 15 to Col 17 Col 17 to Col 9
Observed Calculated Residual (r) Observed Calculated Residual (r) Observed Calculated Residual (r)
difference difference difference difference difference difference
SiO2 1.38 1.34 0.04 2.58 2.57 0.01 0.56 0.50 0.06
TiO 2 0.02 0.08 - 0.06 -0.07 -0.13 0.06 -0.08 -0.06 -0.02
A1203 0.84 0.82 0.02 -0.34 -0.34 0.00 0.54 0.42 0.12
FeO t -0.32 -0.28 -0.04 -0.53 --0.53 0.00 -0.58 -0.64 0.06
MnO 0.00 0.00 0.00 0.00 0.00 0.00 -0.02 -0.01 -0.01
MgO - 1.84 - 1.88 0.04 - 0.95 - 0.95 0.00 -0.47 -0.50 0.03
CaO -0.60 -0.60 0.00 --1.17 -1.17 0.00 0.03 0.12 -0.10
Na20 0.40 0.38 0.02 0.25 0.30 --0.05 0.03 0.12 -0.t0
KzO 0.11 0.14 -0.03 0.24 0.25 -0.01 0.00 0.04 -0.04
St2:0.01 2;r2:0.01 Zr2:0.04
Subtracted Wt.% of Subtracted Wt.% of Subtracted Wt.% of

phases initial phases initial phases initial
magma magma magma
Olivine 3.9 Plagioclase 10.6 Plagioclase 0.1

Clinopyroxene 4.7 Orthopyroxene 1.2 Orthopyroxene 2.1
Plagioclase 4.2 Clinopyroxene 1.5 Titanomagnetite 0.5
Hornblende 7.4
Titanomagnetite 0.5
% Crystallized 12.8 % Crystallized 21.2 % Crystallized 2.7
Trace element calculations Trace element calculations Trace element calculations
low Ko's high KD'S Col 15 low KD's high KD'S Col 17 low KD'S high Ko's Col 9
(Table 2) (Table 2) (Table 2)
Sc 19.4 14.2 18.8 (0.1) 12.3 8.9 14.9 (0.1) a 14.5 13.7 12.7 (0.1)"
V 248 203 179 (21) 135 82 196 (23) a 191 151 150 (25)
Cr 50 15 75 (1) a 18 5 35 (1) a 18 6 25 (1)"
Ni 22 20 51 (8) a 33 20 26 (6) 22 17 10 (7)
Zn 53 43 66 (5) a 49 37 73 (5) a 69 65 70 (4)
Ba 421 414 413 (16) 511 487 497 (17) 510 507 510 (15)
La 10.5 10.4 10.4 (0.5) 12.8 12.5 12.8 (0.5) 13.1 13.0 12.4 (0.5)
Yb 1.89 1.84 1.77 (0.03) a 2.00 1.88 1.96 (0.03) 2.00 1.99 1.83 (0.02) a
Hf 3.1 3.0 3.1 (0.1) 3.7 3.7 3.8 (0.1) 3.9 3.9 3.5 (0.1) a
Numbers in parentheses are counting uncertainties of one standard deviation from I N A A (Table 2)
Indicates unsatisfactory fit
(Sr2). An arbitrary value of s was chosen daughter magma fell between the values calculated
as an upper limit of acceptability. As a further simpli- using the high and low KD'S.
fication, all models calling for the addition of phases Initially, 71 different models were tested to derive
to the parent were rejected. Acceptable fractionation the hornblende-andesites Col 15, 2, 17, 6, and 9 from
models were then subjected to tests involving the trace the olivine-andesite Col 11, and to derive the more
elements Sc, V, Cr, Ni, Zn, Ba, La, Yb, and Hf siliceous hornblende-andesites from the more basic.
as a second stage of the approach. Both surface and 37 of the models passed the major element test, many
total equilibrium conditions were tested (Arth, 1976) with excellent fits, but none of these could successfully
using high and low crystal/liquid distribution coeffi- pass all nine trace element tests under either surface
cients listed in Table 18 and the phase proportions or total equilibrium conditions. Most of the models
calculated from the major element models. A model failed in the prediction of the compatible trace ele-
successfully passed the second test if the concentra- ments (Cr, Ni, and Zn). In each unsuccessful case,
tions of the nine tested trace elements in the presumed the fractionation models predict anomalously low
J.F. Luhr and I.S.E. Carmichael:The ColimaVolcanicComplex,Mexico. I 369
values for these elements. O'Hara (1977) and Walker terminate with a major ashflow eruption in the early
(1979) observed that open system fractionation in- part of the next century. Calculations demonstrate
volving simultaneous magma mixing produces pre- a significant reduction in magmatic water content in
cisely this effect, of anomalous enrichment in highly the most recent, fourth cycle andesites, however, and
compatible elements. The simple fractionation models this may signal a departure from past behavioral pat-
considered here are viewed, then, as evidence in sup- terns.
port of concurrent magma mixing and crystal fraction- Volcfin Colima lies closer to the Middle America
ation in Colima's magma system. Table 25 lists three Trench than any other composite volcano in Mexico,
of the most successful simple fractionation models Colima's post-caldera andesites are depleted in K,
for the generation of compositional trends in the se- Ti, P, and incompatible trace elements (Zn, Rb, Y,
quence Col 13 - Col 15 (1913)- Col 17 (1961) Col 9 Zr, Ba, La, Yb, Hf, Th, and U) (Luhr and Nelson,
(1975). 1979), as may be characteristic of near trench ande-
sites. The investigated suite of post-caldera horn-
blende and olivine-andesites includes products of the
Magma Mixing Models' four major eruptions since 1869, spanning the two
most recent eruptive cycles. From the hornblende-
A major element linear least squares program was andesite ashflow eruption of 1913, which ended the
employed to test the feasibility of forming Col 15, third eruptive cycle, through the fourth cycle horn-
the most basic hornblende-andesite, by mixing Col 11 blende-andesites, there have been increases in whole
with Col 17 or 9, and to test the feasibility of forming rock concentrations of Si, Ba, and Cs, and decreases
Col 17, the next most basic hornblende-andesite, by in Ti, Fe, Mg, Ni, Cr, and Sc. Simple crystal fraction-
mixing Col 15 with Co19. All three models were ation models can closely reproduce major element
rejected on the basis of Zr 2 values greater than 2. variations in the post-caldera suite, but systematically
The magma mixing event which is thought to have fail to predict sufficient concentrations of the compat-
preceded the eruption of Colima's post-caldera oliv- ible trace elements Cr, Ni, and Zn. O'Hara (1977)
ine-andesites is discussed further in Part II of this and Walker (1979) observed that anomalous enrich-
paper. In conclusion, it is not possible to generate all ments of compatible trace elements result when crys-
compositional trends observed within the post-caldera tal fractionation is accompanied by periodic mixing
suite at Volcfin Colima either through simple crystal with relatively basic magma. Simple magma mixing
fractionation or through simple magma mixing using using compositions of investigated andesites as end-
analyzed samples as end members and crystal/liquid members is not a viable mechanism for producing
trace element distribution coefficients listed in Ta- compositional variations within the post-caldera
ble 18. Many crystal fractionation models can closely suite. Analyzed hornblendes in Colima's andesites are
reproduce major element trends of derivative too Mg-rich to sufficiently suppress Fe-enrichment
magmas. Even the best of these models, though, fail in the suite by separation. Small degrees of titanomag-
in the prediction of at least two trace elements. The netite separation are required to derive the observed
systematic failure of simple fractionation models to whole rock trend.
predict sufficient quantities of the compatible trace Pre-eruptive temperature estimates range from
elements Cr, Ni, and Zn, is taken as evidence that 940 ~ 1,000~ C in the hornblende-andesites and 1,030~
Colima's magma system is periodically injected from 1,060~ C in the olivine-andesites. Calculated oxygen
below by pulses of relatively primitive magma. In fugacities lie 1.5 log units above the NNO buffer in
view of the severe restrictions imposed by trace ele- T-fo2 space, and calculated pre-eruptive magmatic
ment data on crystal fractionation models, such mod- water contents range from 0.8-3.6 wt.%.
els which are based solely on major element least Orthopyroxenes and certain clinopyroxenes in all
squares analysis (Dostal etal., 1977; Fairbrothers pre-1961 samples possess relatively Mg-rich rims. The
et al., 1978) should be viewed with caution. most pronounced Mg-rich rims occur in the olivine-
andesites. Analyses of orthopyroxene and olivine phe-
nocryst rim and core compositions in the olivine-
Summary and Conclusions
andesites indicate pre-eruptive mixing involving ande-
Available historical accounts reveal that over the past sitic and basaltic end-member magmas. We emphasize
400 years, Volcfin Colima has evolved through three that magma mixing is not the only process that can
cycles of eruptive activity, each culminating in a volu- produce relatively Mg-rich rims on mafic phenocrysts.
minous ashflow eruption. Andesitic block lava erup- When compared with earlier hornblende-andesites of
tions in 1961 1962 and 1975 1976 initiated activity similar bulk composition, the most recent, fourth cy-
in the fourth historical eruptive cycle which may also cle andesites display: (1) lower modal hornblende
contents, (2) higher t o t a l crystal contents, (3) higher References

c a l c u l a t e d p r e - e r u p t i v e silica activities, a n d (4) lower
c a l c u l a t e d p r e - e r u p t i v e w a t e r contents, in a d d i t i o n to Allen, J.C., Boettcher, A.L., Marland, G. : Amphiboles in andesite
and basalt: I. Stability as a function of P-T-fo 2. Am. Mineral.
the absence of relatively M g - r i c h rims on pyroxenes. 60, 1069-1085 (1975)
All o f these features are consistent with the i n t e r p r e t a - Allen, J.C., Boettcher, A.L.: Amphiboles in andesite and basalt:
tion t h a t the f o u r t h cycle andesites were s o m e w h a t II. Stability as a function of P-T-f~2o-fo 2. Am. Mineral. 63,
less h y d r o u s on e r u p t i o n . It is suggested t h a t the 1074-1087 (1978)
Anderson, A.T.: Magma mixing: Petrological process and vol-
slightly M g - r i c h p y r o x e n e rims in the h o r n b l e n d e -
canological tool. J. Volcanol. Geotherm. Res. 1, 3-33 (1976)
andesites reflect late-stage increases in m a g m a t i c Anderson, A.T., Clayton, R.N., Mayeda, T.K.: Oxygen isotope
w a t e r content, which increase m a g m a t i c f o 2 , and con- thermometry of mafic igneous rocks. J. Geol. 79, 715-729 (1971)
sequently decrease F e + 2 / M g ratios in the m e l t a n d Ando, S. : Minor element geochemistry of the rocks from Mashu
all crystalline phases. The f o u r t h cycle andesites ap- Volcano, Eastern Hokkaido, J. Fac. Sci. Hokkaido Univ. Ser.
4: 16, No. 4, 553-566 (1975)
p a r e n t l y d i d n o t experience a strong, late-stage influx Arreola, J.M. : The recent eruptions of Colima. J. Geol. 11, 749-761
o f water, resulting in lower m a g m a t i c w a t e r contents (1903)
a n d n o r m a l l y z o n e d pyroxenes. Arth, J.G. : Behavior of trace elements during magmatic processes
The a b u n d a n c e o f p y r o c l a s t i c d e p o s i t s a n d h o r n - - a summary of theoretical models and their applications. U.S.
Geol. Survey, J. Res. 4, No. 1, 41M7 (1976)
b l e n d e - b e a r i n g lavas at Volc~n C o l i m a a n d the high
Barcena, M. : Informe Sobre el Estado Actual del Volcan de Co-
fo2 t r e n d o f its m a g m a s reflect the relatively h y d r o u s lima. pp. 40 Mexico: Imp. Sec. Fomento, 1887
n a t u r e o f C o l i m a ' s volcanic system. H i g h w a t e r con- Barker, D.S.: Magmatic trends on alkali-iron-magnesium dia-
tents m a y be a p r i m a r y feature o f C o l i m a ' s n e a r - grams. Am. Mineral. 63, 531-534 (1978)
trench m a g m a s , resulting f r o m v i g o r o u s d e h y d r a t i o n Bender, J.F., Hodges, F.N., Bence, A.E. : Petrogenesis of basalts
from the Project Famous Area: Experimental study from 0
r e a c t i o n s in the u n d e r l y i n g Benioff zone (100 km). to 15 kbars. Earth Planet. Sci. Lett. 41, 277-k02 (1978)
Alternatively, the w a t e r m a y be inherited at n e a r sur- Buddington, A.F., Lindsley, D.H.: Iron-titanium oxide minerals
face levels, either as g r o u n d w a t e r convects into the and synthetic equivalents. J. Petrol. 5, 310-357 (1964)
volcanic system, or as h y d r o t h e r m a l l y altered p o r - Burnham, C.W., Holloway, J.R., Davis, N.F.: Thermodynamic
properties of water to 1000~ C and 1000 Bars. Geol. Soc. Am.
tions of the volcanic pile are assimilated. Slight M g - Spec. Pap. 132, 1-96 (1969)
rich p y r o x e n e rims in C o l i m a ' s pre-1961 h o r n b l e n d e - Carmichael, I.S.E. : The petrology of Thingmuli, a Tertiary volcano
andesites are i n t e r p r e t e d as s u p p o r t i n g a late stage, in eastern Iceland. J. Petrol. 5, part 3, 435-460 (1964)
high level origin for m u c h o f the w a t e r in C o l i m a ' s Carmichael, I.S.E. : The iron-titanium oxides of salic volcanic rocks
magmas. and their associated ferromagnesian silicates. Contrib. Mineral.
Petrol. 14, 36-64 (1967)
A large b o d y of v o l c a n o l o g i c a l a n d p e t r o g r a p h i c Carmichael, I.S.E., Ghiorso, M.S. : Intensive variables in siliceous
evidence n o w exists to s u p p o r t d y n a m i c m o d e l s for magmas. Contrib. Mineral. Petrol. (in press, 1980)
orogenic v o l c a n i c system in which crystal f r a c t i o n a - Carmichael, I.S.E., Nicholls, J. : Iron-titanium oxides and oxygen
tion, m a g m a mixing, a s s i m i l a t i o n o f the v o l c a n i c pile, fugacities in volcanic rocks. J. Geophys. Res. 72, No. 18,
4665-4687 (1967)
a n d o t h e r processes are o p e r a t i n g c o n c u r r e n t l y (Ei- Carmichael, I.S.E., Nicholls, J., Spera, F.J., Wood, B.J., Nelson,
chelberger, 1975; A n d e r s o n , 1976; O ' H a r a , 1977; S.A. : High-temperature properties of silicate liquids: Applica-
F a i r b r o t h e r s et al., 1978 ; J o h n s t o n , 1978 ; Y a n a g i a n d tions to the equilibration and ascent of basic magma. Philos.
Ishizaka, 1978 ; N e l s o n , 1979 ; S a k u y a m a , 1979 ; Stern, Trans. R. Soc. Lond. Ser. A. : 286, 373-431 (1977)
1979). R e c e n t studies o f ocean ridge v o l c a n i s m in- Condie, K.C., Swenson, D.H. : Compositional variation in three
Cascade stratovolcanoes: Jefferson, Ranier, and Shasta, Bull.
creasingly rely on similarly c o m p l e x m o d e l s for ridge Volcanol. 37, No. 2, 205-230 (1973)
m a g m a systems ( D u n g a n and R h o d e s , 1978; R h o d e s Donaldson, C.H., Usselman, T.M., Williams, R.J., Lofgren, G.E. :
et al., 1979; W a l k e r , 1979). C o l i m a ' s olivine-andesites Experimental modeling of the cooling history of Apollo 12 oliv-
r e c o r d p e t r o g r a p h i c evidence o f p r e - e r u p t i v e m a g m a ine basalts. Proceeding Lunar Science Conference 6th, 843-869
(1975)
mixing. In a d d i t i o n , simple crystal f r a c t i o n a t i o n can- Dostal, J., Capedri, S., De Albuquerque, C.A.R.: Calc-alkaline
n o t sufficiently enrich derivative m a g m a s in the c o m - volcanic rocks from N.W. Sardinia: Evaluation of a fractional
patible trace elements Cr, N i , a n d Zn. These o b s e r v a - crystallization model. Bull. Volcanol. 40, No. 4, 1-7 (1977)
tions i m p l y t h a t C o l i m a ' s andesites are also the p r o - Drake, M.J. : Plagioclase-melt equilibria. Geochim. Cosmochim.
Acta 40, 457-465 (1975)
d u c t o f a d y n a m i c s u b v o l c a n i c m a g m a system, fed
Dungan, M.A., Rhodes, J.M. : Residual glasses and melt inclusions
f r o m b e l o w b y pulses o f relatively basic m a g m a . in basalts from DSDP Legs 45 and 46: Evidence for magma
mixing. Contrib. Mineral. Petrol. 67, 417-431 (1978)
Acknowledgements. We thank S.A. Nelson and C.J. Frisch for Eggler, D.H.: Water-saturated and undersaturated melting rela-
assistance in the field, H. Bowman and F. Asaro for NA analyses, tions in a Paricutin andesite and an estimate of water content
and H. Williams for inspiration. Reviews by H. Williams, S.A. in the natural magma. Contrib. Mineral. Petrol. 34, 261571
Nelson, and A. Ewart significantly improved the manuscript. The (1972a)
support of NSF EAR 74-12782 (Carmichael) is gratefully acknowl- Eggler, D.H.: Amphibole stability in H20-undersaturated calc-
edged. alkaline melts. Earth Planet. Sci. Lett. 15, 28-34 (1972b)
Eggler, D.H., Brurnham, C.W.: Crystallization and fractionation canic rocks of Hokkaido, North Japan. J. Fac. Sci. Hokkaido
trends in the system andesite H 2 0 - C O 2 - O 2 at pressures Univ. Ser. 4: 18, No. 3, 449484 (1978)
to 10 Kb. Geol. Soc. Am. Bull. 84, 2517-2532 (1973) Kudo, A.M., Weill, D.F. : An igneous plagioclase geothermometer.
Eichelberger, J.C. : Origin of andesite and dacite: Evidence of mix- Contrib. Mineral. Petrol. 25, 52~55 (1970)
ing at Glass Mountain in California and at other circum- Kuno, H.: Petrological notes on some pyroxene-andesites from
Pacific volcanoes. Geol. Soc. Am. Bull. 86, 1381-1391 (1975) Hakone Volcano, with special reference to some types with
Eichelberger, J.C. : Andesitic volcanism and crustal evolution. Na- pigeonite phenocrysts. Jpn. J. Geol. and Geogr. 13, 107 140
ture 275, No. 7, 21-27 (1978) (1936)
Evans, S.H. Jr., Nash, W.P. : Petrogenesis of xenolith-bearing ba- Kuno, H. : Petrology of Hakone Volcano and the adjacent areas,
salts from southeastern Arizona. Am. Mineral. 64, 249-267 Japan. Geol. Soc. Am. Bull. 61, 957 1020 (1950)
(1979) Kuno, H.: Origin of Cenozoic petrographic provinces of Japan
Ewart, A., Hildreth, W., Carmichael, I.S.E.: Quaternary acid and surrounding areas. Bull. Volcanol. 20, 37 76 (1959)
magma in New Zealand. Contrib. Mineral. Petrol. 51, 1-27 Kyser, T.K., O'Neil, J.R., Carmichael, I.S.E. : Oxygen isotope frac-
(1975) tionations between coexisting phases at high temperatures and
Fairbrothers, G.E., Carr, M.J., Mayfield, D.G.: Temporal mag- their bearing on the genesis of basic lavas and mantle xenoliths.
matic variation at Boqueron Volcano, El Salvador. Contrib. Submitted for publication (1980)
Mineral. Petrol. 67, 1 9 (1978) Lofgren, G.E., Donaldson, C.H., Williams, R.J., Mullins, O. Jr.,
Flowers, G.C. : Correction of Holloway's (1977) adaptation of the Usselman, T.M. : Experimentally reproduced textures and min-
Modified Redlich-Kwong equation of state for calculation of eral chemistry of Apollo 15 quartz-normative basalts. Proc.
the fugacities of molecular species in supercritical fluids of geo- Lunar Sci. Conf. 5th, 549-567, 1974
logic interest. Contrib. Mineral. Petrol. 69, 315-318 (1979) Lofgren, G.E., Usselman, T.M. : Geology, petrology, and crystal-
Foshag, W.F., Gonzalez, R.J. : Birth and development of Paricutin lization of Apollo 15 quartz-normative basalts. Proc. Lunar
Volcano, Mexico, U.S. Geol. Surv. Bull. 965-D, 355489 (1956) Sci. Conf. 6 t~, 79-99, 1975
Francis, P.W., Robool, M.J., Walker, G.P.L., Cobbold, P.R., Cow- Lorenzo, J.L. : Notas Sobre Geologia Glacial del Nevado de Co-
ard, M. : The San Pedro and San Pablo Volcanoes of northern lima. Boletin del Instituto de Geologia, Univ. Natl. Auton.
Chile and their hot avalanche deposits. Geol. Rundsch. 63, Mexico 61, 77-92 (1961)
357-388 (1974) Luhr, J., Carmichael, I.S.E.: Basanite and minette cinder cones
Friedlaender, I. : /~ber die Mexikanischen Vulkane Pico de Ori- from the northern flanks of the andesitic Colima volcanoes,
zabo, Cerro de Tequila, und Colima. Z. Vulkanol. 13, 154-164 Mexico. Geol. Soc. Am. Abstr. Prog. 11, No. 7, 469-470
(1930) (1979)
Fudali, R.F. : Oxygen fugacities of basaltic and andesitic magmas: Luhr, J., Nelson, S.A. : Volcanological and geochemical contrasts
Geochim. Cosmochim. Acta 29, t063-1075 (1965) between two Mexican volcanoes: Colima and Ceboruco. EOS,
Garcia, M.O., Jacobson, S.S. : Crystal clots, amphibole fi'actiona- Trans. Am. Geophys. Union, 1979 PNW Meeting, abstracts
tion, and the evolution of calc-alkaline magmas. Contrib. Min- (1979)
eral. Petrol. 69, No. 4, 319 332 (1979) MacDonald, G.A. Katsura, T. : Eruption of Lassen Peak, Cascade
Gastil, R.G., Jensky, W.: Evidence for strike-slip displacement Range, California, in 1915: Example of mixed magmas. Geol.
beneath the Trans-Mexican Volcanic Belt. Stanford University Soc. Am. Bull. 76, 475482 (1965)
Published in Geol. Sci. 12 13, 171-180 (1973) Mahood, G.A.: A preliminary report on the comenditic dome
Ghiorso, M.S., Carmichael, I.S.E.: A regular solution model for and ashflow complex of Sierra La Primavera, Jalisco. Univ.
met-aluminous silicate liquids: Applications to geothermome- Nal. Auton. Mexico, Inst. Geologia, Revista 1, No. 2, 177-190
try, liquid immiscibility, and the source regions of basic (1977)
magmas. Contrib. Mineral. Petrol. (in press, 1979) Marsh, B.D.: Some Aleutian andesites: Their nature and source.
Green, T.H., Ringwood, A.E. : Genesis of the calc-alkaline igneous J. Geol. 84, 2 7 4 5 (1976)
rock suite. Contrib. Mineral. Petrol. 18, 105-162 (1968) McBirney, A.R.: Petrochemistry of the Cascade andesite vol-
Gutmann, J.T.: Textures and genesis of phenocrysts and mega- canoes. Oregon Dept. of Geol. and Mineral. Ind. 65, 101-107
crysts in basaltic lavas from the Pinacate volcanic field. Am. (1969)
J. Sci. 277, 833-861 (1977) Medaris, L.G., Jr. : Partitioning of Fe z+ and Mg 2+ between coex-
Hatherton, T., Dickenson, W.R. : The relationship between andesi- isting synthetic olivine and orthopyroxene. Am. J. Sci. 267,
tic volcanism and seismicity in Indonesia, the Lesser Antilles, 945 968 (1969)
and other island arcs. J. Geophys. Res. 74, 5301 5310 (1969) Molnar, P., Sykes, L.R. : Tectonics of the Caribbean and Middle
Helgeson, H.C., Delany, J.M., Nesbitt, H.W., Bird, D.K.: Sum- America regions from focaI mechanisms and seismicity. Geol.
mary and critique of the thermodynamic properties of rock Soc. Am. Bull. 80, 1639-1684 (1969)
forming minerals. Am. J. Sci: 278A, 220 pp. (1978) Moorbath, S., Thorpe, R.S., Gibson, I.L. : Strontium isotope evi-
Helz, R.T.: Phase relations of basalts in their melting range at dence for petrogenesis of Mexican andesites. Nature 271,
PH20 = 5 Kb as a function of oxygen fugacity: Part I. Mafic 437439 (1978)
phases. J. Petrol. 14, part 2, 249-302 (1973) Mooser, F. : Active volcanoes of Mexico : International Volc. Assn.
Jakes, P., White, A.J.R.: Major and trace element abundances Catalogue of Active Volcanoes, part VI, 1958
in volcanic rocks of orogenic areas. Geol. Soc. Am. Bull. 83, Mooser, F. : Los Volcanes de Colima. Universidad Nacional, Mex-
29-40 (1972) ico, Geol. Bol. 61, 49 71 (1961)
Johnston, D.A.: Volatiles, magma mixing, and the mechanism Mooser, F. : The Mexican Volcanic Belt - Structure and develop-
of eruption of Augustine Volcano, Alaska. University of Wash- ment. Formation of fractures by differential crustal heating,
ington, Unpublished P h . D . dissert, 177 p. 1978 Pan American Symp. on the Upper Mantle, Mexico, 15-222,
Katsui, Y., Ando, S., Inaba, K. : Formation and magmatic evolu- 1969
tion of Mashu Volcano, East Hokkaido, Japan. J. Fac. Sci. Mooser, F.: The Mexican Volcanic Belt: Structure and tectonics.
Hokkaido Univ. Ser. 4: 16, No. 4, 533-552 (1975) Geofis. Int. 12, 55-70 (1972)
Katsui, Y., Oba, Y., Ando, S., Nishimura, S., Masuda, Y., Kura- Mooser, F., Maldonando-Koerdell, M.: State of the volcanoes
sawa, H., Fujimaki, H. : Petrochemistry of the Quaternary vol- Colima Volcano, renewed activity. Mexican National Report
on Volcanology, XI th General Assembly of the I.U.G.G., Ex- Stormer, J.C., Jr., Carmichael, I.S.E. : The Kudo-Weill plagioclase
trait du Bull. Volcanol. ser. 2, Tome 21, 164 (1959) geothermometer and porphyritic acid glasses. Contrib. Mineral.
Mooser, F., Maldonando-Koerdell, M.: State of the volcanoes Petrol. 28, 306-309 (1970)
- Colima Volcano. Mexican National Report on Volcanology, Sugimura, A. : Spatial relations of basaltic magmas in island arcs.
XIIP b General Assembly of the I.U.G.G., 1963 In: Treatise on Basalts. (H.H. Hess and A. Poldervaart eds.),
Mysen, B.: Partitioning of iron and magnesium between crystals pp. 537-572. Vol. 2. INTERSCIENCE, New York (1968)
and partial melts in peridotite upper mantle. Contrib. Mineral. Taylor, S.R. : Trace element chemistry of andesites and associated
Petrol. 52, 69-76 (1975) calcalkaline rocks. Oregon Dept. Geol. and Miner. Ind. Bull.
Nakamura, Y.: Geology and petrology of Bandai and Nekoma 65, 43-63 (1969)
volcanoes. Sci. Rep. Tohoku Univ. Ser. 3, 14, No. 1, 67-119 Thorpe, R.S. : Tectonic significance of alkaline volcanism in eastern
(1978) Mexico. Tectonophysics 40, 19~6 (1977)
Nelson, S.A.: The geology and petrology of Volc~in Ceboruco, Thorpe, R.S., Gibson, I.L., Vizcaino, J.S.: Andesitic pyroclastic
Nayarit, Mexico. University of California, Berkeley, Unpub- flows from Volcfin Colima. Nature 265, 724-725 (1977)
lished Ph.D. dissert., Part I, 164 p. 1979 Ulmer, G.C., Rosenhauer, M., Woermann, E., Ginder, J., Drory-
Newhall, C.G. : Temporal variation in the lavas of Mayon Volcano, Wolff, A., Wasilewski, P. : Applicability of electrochemical oxy-
Philippines. Geol. Soc. Am. Abstr. Prog. 10, No. 3, 139 (1978) gen fugacity measurements to geothermometry. Am. Mineral.
O'Hara, M.J. : Geochemical evolution during fractional crystalliza- 61, 653-660 (1976)
tion of a periodically refilled magma chamber. Nature 266, Wager, L R . : The major element variation of the layered series
No. 7, 503-507 (1977) of the Skaergaard intrusion and a re-estimation of the average
Osborn, E.F. : Reaction series for subalkaline igneous rocks based composition of the hidden layered series and of the successive
on different oxygen pressure conditions. Am. Mineral. 47, residual magmas. J. Petrol. 1, 364-398 (1960)
211-226 (1962) Walker, D. : Phase equilibrium curvature effects on mixed magma.
Peacock, M.A.: Classification of igneous rock series. J. Geol. 39, EOS. Trans. Am. Geophys. Union, 1979 PNW Meeting, abstr.,
54-67 (1931) 1979
Perlman, I., Asaro, F.: Pottery analysis by neutron activation. Waitz, P.: Le Votcan de Colima. 10th International Geological
Archaeometry 11, 21-52 (1969) Congress, Field Trip Guide, chap. 13, p. 27, 1906
Pichler, H., Weyl, R. : Quaternary alkaline rocks in eastern Mexico Waitz, P. : Der Gegenw/irtige Zustand der Mexikanischen Vulkane
and Central America. Munster. Forsch. Geol. Palaont. 38-39, und die Letzte Eruption des Vulkans von Colima (1913). Z.
159 178 (1976) Vulkanol. 1, 247 (1915)
Powell, R. : Thermodynamics of coexisting cummingtonite-horn- Waitz, P.: Datos historicos y bibliograficos arcerca del Volcfin
blende pairs. Contrib. Mineral. Petrol. ill, 29-37 (1975) de Colima. In: Mem. Soc. Cient (Antonio Alzate, ed.), Vol.
Rhodes, J.M., Dungan, M.A., Blanchard, D.P., Long, P.E.: 53, pp. 349-384 Mexico: Tomo 1932
Magma mixing at mid-ocean ridges: Evidence from basalts Wells, P.R.A. : Pyroxene thermometry in simple and complex sys-
drilled near 22~ N on the Mid-Atlantic ridge. Tectonophysics tems. Contrib. Mineral. Petrol. 62, 129 139 (1977)
55, 35~1 (1979) Weyl, R. : Vulkanismus und Vulkanlandschaft im Hochland yon
Rittmann, A. : Volcanoes and their activity: Trans. of 2 na German Mexiko. Nat. Mus. 104, No. 5, 137-152 (1974)
Edition by E.A. Vincent, INTERSCIENCE Publ., New York. White, S.E. : A geologic investigation of the late Pleistocene histroy
305 pp. (1962) of the volcano Popocatepetl, Mexico. Syracuse University, Un-
Robin, C., Tournon, J. : Spatial relations of andesitic and alkaline published Ph.D. dissert., 350 p. 1952
provinces in Mexico and Central America. Can. J. Earth Sci. Wilcox, R.E. : Petrology of Paricutin Volcano, Mexico. U.S. Geol.
15, 1633 1641 (1978) Surv. Bull. 965-C, 281-349 (1954)
Roeder, P.L., Emslie, R.F. : Olivine-liquid equilibrium. Contrib. Williams, H. : Volcanoes of the Paricutin Region. U.S. Geol. Surv.
Mineral. Petrol. 29, 275~89 (1970) Bull. 965-D, 165-275 (1950)
Rose, W.I., Jr., Grant, N.K., Hahn, G.A., Lange, I.M., Powell, Williams, R.J., Eugster, H.P. : An experimental study of (Fe, Mg),
J.L., Easter, J., DeGraff, J.M. : The evolution of Santa Maria olivine (Fe, Mg) pyroxene reactions and their geological
Volcano, Guatemala. J. Geol. 85, 63-87 (1977) applications. Geol. Soc. Am. Abstr. Prog. 237, (1969)
Sakuyama, M. : Evidence of magma mixing: Petrological study Wood, B.J., Banno, S. : Garnet-orthopyroxene and orthopyroxene-
of Shirouma-Oike calc-alkaline andesite volcano, Jpn. J. Volca- clinopyroxene relationships in simple and complex systems.
nol. Geotherm. Res. 5, 179-208 (1979) Contrib. Mineral. Petrol. 42, 109-124 (1973)
Sapper, K.: VULKANKUNDE. pp. 424 J. Engelhorns Nachf. Woodruff, L.G., Anderson, A.T.: An arrested state of magma
Stuttgart: 1927 mixing, Pavlof Volcano, Alaska: EOS, Trans. Am. Geophys.
Sartorius, C.: Eruption of the volcano of Colima in June, 1869. Union, 1979 PNW Meeting, abstr. 1979
Am. J. Sci. 102, 381-383 (1871) Wright, T.L.: Presentation and interpretation of chemical data
Sperry, F.L.: The eruption of Colima. Am. J. Sci. 15, 487-488 for igneous rocks. Contrib. Mineral. Petrol. 48, 233 248 (1974)
(1903) Wright, T i . , Doherty, P.C.: A linear programming and least
Stern, C.R., Huang, W., Wyllie, P.J.: Basalt-andesite-rhyolite- squares computer method for solving petrologic mixing prob-
H20: Crystallization intervals with excess H20 and H20-un- lems. Geol. Soc. Am. Bull. 81, 1995-2008 (1970)
dersaturated liquidus surfaces to 35 kbars, with implications Wright, T.L., Weiblen, P.W. : Mineral compositions and paragen-
for magma genesis. Earth Planet. Sci. Lett. 28, 189-196 (1975) esis in tholeiitic basalt from Makuopuhi lava lake, Hawaii.
Stern, R.J.: On the origin of andesite in the Northern Mariana Geol. Soc. Am. Spec. Pap. 115, 242~43 (1968)
Island Arc: Implications from Agrigan. Contrib. Mineral. Pet- Yanugi, T., Ishizaka, K. : Batch fractionation model for the evolu-
rol. 68, No. 2, 207-219 (1979) tion of volcanic rocks in an island arc: An example from central
Stewart, D.C. : Crystal clots in calc-alkaline andesites as breakdown Japan. Earth Planet. Sci. Lett. 40, 252 262 (1978)
products of high-A1 amphiboles. Contrib. Mineral. Petrol. 53,
195-204 (1975)
Stoiber, R.E. : Products of the fumaroles, Colima Volcano, Mexico.
Yearb. Am. Philos. Soc. 295-297 (1967) Received September 5, 1979; Accepted November 1, 1979

Luhr 1980

Uploaded by

Copyright:

Available Formats

Luhr 1980

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Luhr 1980

Uploaded by

Copyright:

Available Formats

Contributions to

Contrib. Mineral. Petrol. 71, 343-372 (1980) Mineralogy and

The Colima Volcanic Complex, Mexico

James F. Luhr and Ian S.E. Carmichael

Fig. 2. Generalized map of the

ERUPTIVE PERIODICITY AT VOLCAN COLIMA

CYCLE 1 ~]4 CYCLE 2 :!.CYCLE 3=!. CYCLE 4

1500 1600 1700 1800 1900 2000

PRECEDING ASHFLOW -- I•! 1818 1913

Table 1. Whole rock wet chemical analyses

Analyses by I.S.E. Carmichael

Table 2. Neutron activation analyses of whole rock samples a

Na% 3.66 3.56 3.64 3.35 3.67 3.68 3.13 3.15

Eu/Eu *b 1.01 1.06 0.95 1.01 1.01 1.00 1.02 1.04

In ppm, except Na, Ti, and Fe

Col 9 6 17 15 18 Col 9G 6G 17G 15G 2G 7G llG

Cu 27 35 30 23 18 Na% 3.94 4.07 4.16 3.94 4.16 3.90 3.90

Table 6. Modal analyses (vol.%) ( > 700 points counted)

9ph 1975 30.4 1.8 1.9 - 0.8 tr 52.8 47.2

ph: phenocrysts ( > 0.3 m m ) ; mp : microphenocrysts ( > 0.03 m m )

Table 7. Microprobe analyses of plagioclase

Col. 2r 2ab 2an 7r 7ab 7an 11 r 1 tab 11 an 13r 13 ab 13an

Table 8. Neutron activation analyses of plagioclase" P'~CN~ t ~ ' P -- ~D

Na% 4.04 4.00 4.00 3.78 3.92 4.08 3.53

Na% 0.21 0.83 0.48 0.19 0.12 0.I3

Table 11. Average microprobe analyses of clinopyroxene

Col 9r 9c 6 17r 17c 15r 15c 18 2 7 11r 11c 13r 13c

r: average of phenocryst rims; c: average of phenocryst cores

Col 6 17 t5 18 2 7 Col 6h 17h 15h 2h 7h 11 ol

Table 15. A v e r a g e m i c r o p r o b e analyses of olivine

Col 9x 17x llr llc llx 13r 13c 13x

SiO2 37.94 39.00 38.14 38.40 39.25 38.53 39.15 40.60

Total 100.10 100.27 100.52 100.68 98.83 100.80 101.30 101.37

Xfo 0.739 0.809 0.733 0.751 0.878 0.761 0.794 0.871

afo and afa calculated after Carmichael et al. (1977)

N o reliable analysis of phenocrystic hornblende was obtained Olivine

Table 16. Average microprobe analyses of F e - Ti oxides

Col 9m 9h 6m 17m 15m 18m 2m 7m llm ~ 11m b t3m ~ 13m b

Table 18. Ranges of weight fraction distribution coefficients

plag opx cpx hbd oliv titanmt

Sc 0.01 - 0.03 2.2 4.7 6.9 16 8.1 13 0.30 1.7 - 3.3

Distribution Coefficients bly precipitate in such subvolcanic magma chambers

Estimation of Intensive Variables

Table 20. Temperature estimations for Colima's post-caldera andesites (~ C)

9 940 986 992 968 882 - - 880 1,020 1,021 (2.2)

15 1,000 - 1,023 1,010 1,028 - - 1,200 1,200 1,192 (0.2)

2 940 - 971 963 . . . . . 1,052 (2.5)

Ta:Assumed pre-eruptive temperature T6: O18/O16 cpx-oliv (Kyser et al., 1980)

equilibration of phenocryst rims of 1,120 ~ C, 1 kb, Oxygen Fugacity

1 6MgzSiO4 + 6FeSiO3 + 02 = 2Fe304 + 12MgSiO3 27,081 -8.874 -0.20

Fig. 7. T'-fo2 plot including data points for

Eq. 4 Ghiorso and Carmichael (1980) Col 9 6 17 15 2 7 11

Table 24. Calculation of magmatic water content

Col T~ T~ P(bars) 7.hba

melt melt melt w t . oY0H20

Composite Phenocryst Textures and

OLIV OPX CPX PLAG

quenched scoria of Col 15, however, whose horn-

percentage of glass inclusions in plagioclase generally r*o*r~%

percentage of glass inclusions in plagioclase generally ror~%