USGS Logo Geological Survey Circular 838
Guides to Some Volcanic Terrances in Washington, Idaho, Oregon, and Northern California


Stanley A. Mertzman
Department of Geology, Franklin and Marshall College, Lancaster, PA 17604


The Medicine Lake Highland is an ares of diverse volcanism whose extrusives span the last 1.5 to 2 million years of earth history. The enigmatic Andesite Tuff of C. A. Anderson (1941) has been provisionally placed into the general stratigraphic sequence as one unit within the collage collectively known as the "pre-Medicine Lake andesite" volcanics. This proposed positioning results from field, petrographic, and chemical evidence which suggests the existence of m series of andesitic lava flows that are coeval in age or slightly younger than the Andesite Tuff. One of these equivalent lava flows has given an age of 0.21±0.05 m.y. old. If this correlation is accurate, then the conclusion that the Andesite Tuff represents an event which occurred prior to the initial eruptions of the Medicine Lake volcano is true and, therefore, the tuff cannot be correlated with the development of the Medicine Lake summit caldera.

Additional silicic volcanism, including rhyolitic domes and flows, is considerably older (0.43 to 0.95 m.y.); a fact which indicates no direct genetic link need exist between the rhyolites and the tuff. Lastly, the rhyolites can be divided into two separate groups on the basis of SiO2, neither of which overlaps that of the Holocene glass flows, a feature for which the Medicine Lake Highland is well known.


Several recent summaries by Macdonald (1972) and Williams and McBirney (1979) concerning the origin of collapse calderas suggest a number of these features formed concurrently with the eruption of ash-flow tuffs. Noble (1969) suggested that such a relationship may in fact explain the origin of the Medicine Lake caldera in northern California. In particular he proposed that a unit known as the Andesite Tuff may correlate with the collapse of the Medicine Lake caldera and, therefore, post-date the extrusion of the "older shield-forming andesite" unit of C. A. Anderson (1941). This interpretation was also reiterated by A. T. Anderson (1976). On the other hand, Mertzman (1977) interpreted these olivine andesite lavas as the initial eruptions of the Medicine Lake volcano and therefore necessarily pre-caldera in age, a conclusion similar to that of C. A. Anderson (1941). These contradictory interpretations are abetted by a lack of outcrop in critical areas and the inability to obtain a reliable K/Ar age date from the Andesite Tuff, a difficulty arising from its large scale atmospheric argon contamination. However, field mapping by Walter (1975) and Mertzman (unpublished data) indicate the existence of andesitic lavas which appear to be coeval with the andesitic ash-flow tuff unit. Subsequent analytical studies concerning these tuff-equivalent lavas has apparently resolved the contradictory interpretations previously outlined and provides a rudimentary stratigraphy on which to build future research in the Medicine Lake Highland.


The Medicine Lake Highland, located approximately 50 km east-northeast of Mount Shasta in northern California, is the southern extension of a discontinuous belt of shield-like composite volcanoes which lie 30-50 km east of the High Cascade stratovolcanoes. The northern extension of this belt includes Newberry volcano in central Oregon and Simcoe volcano in south-central Washington. Previous research in the Medicine Lake Highland has been recently summarized by Mertzman (1977) and Heiken (1978), and so the writer will present only the highlights of the geologic history. Over the past several years three dozen K-Ar analyses have been performed as part of an ongoing attempt to ascertain the geochronologic development of the region surrounding the Medicine Lake Highland and extending toward Mount Shasta. To date, all the samples from within the region covered by Figure 1 are less than two million years old. The oldest lavas are the basaltic rocks of Timber Mountain volcano, which is located just off Figure 1 approxiately 9 miles southeast of Lava Beds National Monument. These lavas range from 1.7 to 1.9 million years old. Thin flows of low potassium high-alumina olivine tholeiite are well exposed along the portion of Gillem's Bluff fault scarp within Lava Beds National Monument and vary from 1.0 to 1.4 m.y. old. In addition, numerous olivine tholeiite flows of nearly identical petrography and chemistry outcrop both around the periphery and intercalated with lavas of the Medicine Lake Highland and range in age from approximately 1.4 m.y. old to Late Pleistocene to Recent. Moving from the Highland westward there are approximately a dozen mostly andesitic composite volcanoes within the Dray, Dorris and Mount Dome quadrangles whose ages fall between 0.3 and 1.3 m.y. old. It is within the context of this volcanic history that the silicic igneous activity on the northern and western margins of the Highland must be viewed and its origin eventually explained.

Figure l. Geologic map of selected rock units from the Medicine Lake Highland area. The stratigraphic column is based in part on K/Ar dating. (click on image for an enlargement in a new window)


K-Ar results

The K/Ar data are shown in Table 1 with the exact sample locations depicted in Figure l. Potassium was determined together with all the other major elements utilizing a lithium tetraborate fusion which results in a one inch glass disc. The actual analysis was performed using an automated vacuum X-ray fluorescence spectrometer. Argon was determined by pipette-spiking with 38Ar; the isotopic measurements performed with a MS-10 mass spectrometer. This instrument displays virtually no memory and no mass 36 background. The argon fusion was carried out on 3-6 grams of 8-42 mesh fragments of whole rock which had been ultrasonically cleaned in distilled water and alcohol, dried in an oven and pre-baked in a vacuum at 250°C. The uncertainty in the age is estimated utilizing a 0.3% uncertainty in the argon spike, a 2% uncertainty in sample homogeneity, and an absolute uncertainty in the potassium of 0.01% K2O.

TABLE l. Analytical results of K-Ar dating.a

Map Location
40Arb/g 40Arb/40Ar
X 100
K-Ar age

Rhyolite domes
ML4918 0.026911.62 4.420.43±0.04
SM5117 0.041268.04 4.630.61±0.03
56B115 0.04435.55 3.180.95±0.14
Lava flow equivalent of Andesite Tuff
263C112 0.00653.18 2.140.21±0.05
Earliest Medicine Lake volcano lavas
366WA 0.001152.12 0.850.09±0.05
SM25cB 0.000970.74 0.850.08±0.04

aConstants used are: 40K/total K = 1.167 x 10-4; λβ = 4.962 x 10-10yr-1; λβ = 5.81 x 10-11yr-1.

bRadiogenic argon, 10-10 moles.

cThe age is slightly different from that reported in Mertzman (1977) as a result of recalculation with the new K-Ar age constants of Dalrymple and Mankinen (1979).

The silicic rocks dated in this investigation are from rhyolitic flows and domes. The contact relationships between the domes and the surrounding volcanics generally provide rather inconclusive information. These rhyolites range from 0.43 to 0.95 m.y. old with relatively small uncertainties. This spread of ages falls within the range previously quoted for the composite volcanoes located generally west of the Medicine Lake Highland. Any theory which attempts to explain the petrogenesis of this region over the past two million years must come to grips with this space-time-composition relationship which exists between these two groups of volcanic rocks.

The Andesite Tuff is the one major ignimbritic unit in the region and its areal distribution is outlined in figure l. In general, it is a poorly welded ash-flow tuff that completely lacks any stratification. However, where the base is exposed (e.g. in the southern portion of the area now known as the "Bighorn Sheep Enclosure" on Gillem Bluff) substantial flattening of pumice lapilli has occurred together with significantly more welding. The writer has made three attempts to radiometrically date this unit using a moderately welded sample, a poorly welded sample, and a plagioclase separate (no potassium feldspar present). None of these attempts were successful due to large scale atmospheric argon contamination. Subsequent detailed field work in the area south of Dock Well by Walter (1975) and myself (See portion of figure 1 delineated as "Flows equivalent of Andesite Tuff") led to the isolation of several andesitic flows which appeared to be the same age or slightly younger than the Andesite Tuff. Petrographic and chemical studies supported the mapping interpretation which has been further substantiated by reconnaissance microprobe mineral analysis. These data indicate that smooth compositional gradations as well as a large measure of overlapping occur between the proposed equivalent lava flows and the tuff with respect to the orthopyroxene, clinopyroxene and plagioclase feldspar. In fact, the plagioclase phenocrysts which occur in both the flows and the tuff have virtually the same zoning pattern in addition to highly similar chemistries. One of the proposed tuff-equivalent lava flows gives an age of 0.21±0.05 (Table 1). I submit that this age closely reflects the time of eruption of the Andesite Tuff and, at the very worst, reflects its minimum age.

The Andesite Tuff controls the topography on the northwest side of the Medicine Lake Highland onto which the earliest of the Medicine Lake basaltic andesite flows were extruded. Two stratigraphically equivalent lava flows, from the northwest and the southwest flanks of the Medicine Lake volcano (B and A respectively in figure 1) have given ages of 0.08±0.04 and 0.09±0.05 m.y. old. Since these flows represent both the initial eruptions of the Medicine Lake volcano as well as the bulk of the extrusives which compose the Highland (equivalent to the shield-forming platy olivine andesite of C. A. Anderson, 1941) the conclusion can be drawn that the entire 900-1000 m thickness of lavas related to the Medicine Lake volcano (See Mertzman, 1977) has accumulated over the past 100,000 years. A corollary of this conclusion is that the caldera-forming event must be <100,000 years old; unfortunately no minimum age can yet be assigned since no rampart-forming andesite, which represents syn-to post-caldera volcanism, has heretofore been radiometrically dated - a void which will be filled over the next several months. Two additional points are worthy of note: since the stratigraphic thickness of the tuff noticeably increases in a southeasterly direction and yet it does not outcrop on the eastern and southern margins of the Highland, it seems safe in assuming that the vent area for the tuff lies buried beneath the younger andesitic lavas of the Medicine Lake volcano. Secondly, the gross age distinctions which exist between the rhyolitic domes and flows and the tuff-equivalent andesitic lavas on the northwest margin of the Highland, which differ by a factor of 3 to 4, indicate separate pulses of volcanic activity rather than the rhyolite and the tuff being time correlative. The latter interpretation was suggested as a possibility by C. A. Anderson (1941).


Major and trace element chemistry was performed by X-ray fluorescence spectrometry and the resulting data together with CIPW norms are reported in Tables 2 and 3; the exact sample locations are depincted in Figure 1. For trace element analysis one gram of dried whole rock powder was mixed with 0.5 g of microcrystalline cellulose and pressed into pellets using a course grade of cellulose as a backing. Trace element data reduction was achieved by the mass absorption correction method outlined by Hower (1959).

TABLE 2. Major and trace element chemistry together with C.I.P.W. norms of the Andesite Tuff.
The exact sample locations are depicted in figure 1.

Map location number
Sample number

SiO2 59.2660.4560.95 61.0261.0661.29 61.4665.69
Al2O3 18.2415.7317.00 15.9416.6615.77 15.2714.93
FeO 3.100.911.69 1.421.921.41 1.461.02
Fe2O3 3.284.884.36 4.533.564.26 3.882.65
TiO2 1.030.900.95 0.920.900.90 0.860.52
MnO 0.100.10
MgO 1.981.861.90 1.891.742.28 1.651.11
CaO 3.834.083.71 4.193.944.08 3.762.67
Na2O 4.554.274.67 4.604.724.79 4.714.35
K2O 1.532.282.17 2.282.362.82 2.462.73
P2O5 0.200.13
Total 100.4098.99100.41 100.29100.07100.00 98.79100.42

Q 16.1717.0815.58 15.4814.7312.84 16.2623.90
C 2.74-0.69 --- -0.28
Or 9.3114.0813.12 13.8814.3516.98 15.1716.82
Ab 39.6337.7740.43 40.0841.1041.30 41.6038.38
An 18.0917.7917.57 16.5917.8113.45 13.8412.94
Di -1.57- 2.710.713.43 3.33-
Hy 6.634.114.84 3.594.134.20 2.742.88
MT 4.900.713.22 2.344.022.34 2.652.20
Ilm 2.011.791.85 1.801.761.74 1.701.03
Hm -4.612.24 3.050.892.73 2.221.25
Ap 0.530.490.45 0.480.500.99 0.490.32

Rb 305446 545172 6468
Sr 373389373 387394377 360293
Ni 626 442 65
Zr 275262260 255263242 273213
Ba 570552640 588622523 571656
Y 332623 172431 2125
V 11510199 12110672 9430
Rb/Sr 0.0800.1390.123 0.1400.1290.191 0.1780.232
K/Rb 423351392 351384325 319333

*Loss on ignition

TABLE 3. Major and trace element chemistry together with C.I.P.W. norms of the lava flow equivalent of the Andesite Tuff (9-12). Also included are six analyses (13-18) of older silicic volcanic rocks.
The exact sample locations are depicted in figure l.

Map location number
Sample number

SiO2 59.5560.5060.5960.8471.08 71.3071.4771.9971.96 76.38
Al2O3 16.5016.7116.4516.9814.25 14.3214.4814.1112.93 12.13
FeO 4.361.883.744.560.70 1.070.910.600.82 0.70
Fe2O3 1.434.721.681.411.78 1.781.632.230.42 0.38
TiO2 0.740.910.690.790.32 0.440.310.450.21 0.05
MnO 0.04
MgO 3.162.662.893.450.53 0.420.510.340.08 <.02
CaO 5.564.885.255.741.74 1.921.661.240.86 0.46
Na2O 3.584.183.693.854.73 5.104.784.874.06 4.46
K2O 4.42
P2O5 0.01
Total 98.70100.2498.84100.8399.45 100.5199.95100.6099.34 99.47

Q 13.3015.1514.0911.59 27.4826.0927.6824.6331.81 32.92
C ---- 0.04-0.37--
Or 12.4012.1514.2812.63 19.1417.9318.9725.2327.62 26.38
Ab 31.1635.8131.9832.54 40.6543.3440.8241.1134.68 38.11
An 23.5921.1021.8722.71 8.247.307.723.993.42 0.03
Di 2.891.792.893.43 -1.21-1.270.55 1.83
Hy 12.635.8810.6013.01 1.340.491.280.260.81 0.07
MT 2.133.862.492.04 1.582.442.250.790.61 0.56
Ilm 1.451.751.341.50 0.620.840.590.850.40 0.10
Hm -2.12-- 0.720.100.091.68- -
Ap 0.450.400.450.56 0.02


*Loss on ignition
ND=not detected

Analyses of the Andesite Tuff are found in Table 2; typically these samples have relatively large loss on ignition values, ranging from 1.87 to 4.52% by weight, a point which encouraged the writer to calculate the CIPW norms on the basis of recalculated analyses derived by normalizing to 100 percent anhydrous totals. Cast in this light virtually all the tuff samples become more dacitic than andesitic in bulk composition. The only additional facet of their chemistry worthy of highlighting at this time is that three of eight samples are peraluminous. Comparison of the chemistry of the tuffs with that of the four samples for the proposed equivalent andesitic lava flows yields a pattern of lower SiO2 and Na2O and higher MgO and CaO for the lavas. Also, the percentage of phenocrysts is higher in the lavas than in the tuff samples (See Table 4). These trends are analogous to those reported from numerous other ignimbritic areas (e.g., Lipman and Others, 1966; Ratte and Steven, 1967; Christiansen, 1979), further substantiating this writer's claim of a co-genetic relationship linking the lavas and the andesitic tuff.

TABLE 4 Modal analyses (volume percent). The data reported are average values (8 and 4 samples, respectively) which cover the entire range of petrographic variation within both the Andesite Tuff and the lava flows proposed as being equivalent to the tuff.

MineralAndesite TuffTuff equivalent lavas


The rhyolitic volcanics form two distinct groups on the basis of their SiO2 contents: a low SiO2group (71-72%) and a high SiO2 group (>74.5%). Examining the age and chemical data concurrently produces a noteworthy correlation between decreasing age and increasing SiO2 content. Further documentation of this emerging pattern is presently underway. Whether or not this age-composition trend can be genetically linked to a viable crystal fractionation model awaits mineral composition data concerning phenocryst-forming phases, thus enabling some quantitative modelling to be performed. It is also intriguing to note that the chemistry of these older rhyolitic extrusives does not mimic that of the Recent silicic glass flow activity in the Medicine Lake Highland (Eichelberger, 1975; Heiken, 1978; Mertzman, in press). Table 5 presents average chemical analyses for the three Recent glass flows. It is evident that these data form a third group, intermediate between the two groups defined by the older silicic volcanism. Sufficient chemistry is currently available, all performed with the same analytical procedures, to probably eliminate insufficient sampling and analytical uncertainty as potential factors producing the clump-like patterns rather than a continuum of elemental variation. Whether or not this older episode (or episodes) of silicic volcanism is monogenetic or polygenetic, is a research problem currently under study utilizing microprobe and isotopic techniques.

TABLE 5.Major and trace element analyses of three Holocene glass flows from the Medicine Lake Highland. (n = number of samples analyzed).

n Little
Glass flow

SiO2 73.4173.5773.23
Al2O3 13.6213.6813.56


*Loss on ignition.
ND = not detected.


Analytical data concerning the older silicic magmatic activity on the northern flanks of Medicine Lake Highland is presented. Available K/Ar results give a 0.43 to 0.95 m.y. range for the rhyolitic volcanism, which can be divided chemically into low silica (71-72%) and high silica (>74.5%) groups. Both of these groups are clearly distinguishable from the Holocene glass flows with respect to chemistry. The areal extent and chemical variability of the Andesite Tuff is outlined. The past inability to radiometrically date the time of eruption of the tuff has tentatively been circumvented by analyzing one of several lava flows which were erupted concurrently with or very shortly after the tuff. The age of the equivalent flow is 0.21±0.05 m.y. old, a date that clearly establishes the eruption of the tuff as an event preceding the development of the Medicine Lake andesitic volcano and, therefore, not correlative with the formation of its summit caldera.


I thank the Research Corporation for a Cottrell College Science Grant, the Franklin and Marshall committee on Grants, and the American Philosophical Society for financial support of the field work. I also thank Dr. James Aronson of Case Western Reserve University for access to his K/Ar laboratory as well as for his timely advice. The automated vacuum x-ray fluorescence spectrometer was purchased with grants from the Pennsylvania Science and Engineering Foundation, the National Science Foundation, the Gulf Oil Company, and the Fleischmann Foundation.


Anderson, A. T., 1976, Magma mixing, petrological process and volcanological tool; Journal of Volcanology and Geothermal Research, v. 1. p. 3-33.

Anderson, C. A., 1941, Volcanoes of the Medicine Lake Highland, California: University of California Publications, Bulletin of the Department of Geological Sciences, v. 25, no. 7, p. 347-422.

Christiansen, R. L., 1979, Cooling units and composite sheets in relation to caldera structure: Geological Society of America Special Paper 180.

Dalrymple, G. B., and Mankinen, E. A., 1979, Revised geomagnetic polarity time scale for the interval 0-5 m.y. B. P.: Journal of Geophysical Research, v. 84, p. 615-626.

Eichelberger, J. C., 1975, Origin of andesite and dacite: evidence of mixing at Glass Mountain in California and at other circum-Pacific volcanoes: Geological Society of America Bulletin, v. 86, p. 1381-139-1.

Heiken, G., 1978, Plinian-type eruptions in the Medicine Lake Highland, California, and the nature of the underlying magma: Journal of Volcanology and Geothermal Research, v. 4, p. 375-402.

Hower, J., 1959, Matrix corrections in the x-ray spectrographic trace element analysis of rocks and minerals: American Mineralogist, v. 44, p. 19-32.

Lipman, P. W. Christiansen, R. L., and O'Connor, J. T., 1966, A compositionally zoned ash-flow sheet in southern Nevada: U.S. Geological Survey Professional Paper 524-F, 47 p.

Macdonald, G. A., 1972, Volcanoes, Prentice-Hall, New Jersey, 510 p.

Mertzman, S. A., 1977, The petrology and geochemistry of the Medicine Lake volcano, California: Contributions to Mineralogy and Petrology, v. 62, p. 221-247.

Mertzman, S. A., in press, The petrogenesis of Recent silicic magmatism in the Medicine Lake Highland: evidence from cognate inclusions found at Little Glass Mountain, California: Geochimica et Cosmochimica Acta.

Noble, D. C., 1979, Speculations on the origin of the Medicine Lake caldera: Oregon Department of Geology and Mineral Industries Bulletin 65, p. 193.

Ratte J. C., and Steven T. A., 1967, Ash flows and related volcanic rocks associated with the Creede caldera San Juan Mountains, Colorado: U. S. Geological Survey Professional Paper, 524-H, 58 p.

Walter, R. C., 1975, Geology and petrology of the northwest portion of the Medicine Lake Highland, California: Unpub. B. A. thesis, Franklin and Marshall College.

Williams, H. and McBirney A. B., 1979, Volcanology: Freeman, Cooper and Co., San Francisco, 397 p.

<<< Previous <<< Contents >>> Next >>>

Last Updated: 28-Mar-2006