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



Rocks of the John Day Formation form three distinct mineralogical and chemical groups, Tuffaceous claystones and vitric tuffs are chiefly andesitic to dacitic, ash-flow tuffs and minor silicic lava flows are rhyolitic, and mafic lava flows are alkali-olivine basalt or trachyandesite.

Tuffaceous claystone and vitric tuff - The original composition of the tuffaceous claystone and vitric tuff cannot be determined directly by chemical analysis. Much of the original glassy material in these rocks has been replaced by montmorillonite or clinoptilolite during weathering before burial or by later diagenetic alteration. Chemical analyses of altered rocks reported by Hay (1963) indicate extreme hydration accompanied by leaching of Na2O. Chemical analsyes of glassy material from the upper part of the formation are equally misleading because of extreme etching, bleaching, and hydration of the glass. The initial composition of the tuffaceous claystone and vitric tuff must therefore be inferred from the pyrogenic mineralogy. Crystals in these rocks are chiefly andesine feldspar with lesser amounts of labradorite, magnetite, pyroxene, hornblende, biotite, and ilmenite suggesting that the bulk of the material was originally andesitic to dacitic in composition (Table 1; Hay, 1962a; 1963; Fisher and Rensberger, 1972). Less abundant oligoclase-bearing tuff and tuffaceous claystone was probably rhyodacitic and a few tuffs in the lower part of the formation containing quartz, sanidine, and oligoclase were probably rhyolitic in compostion. These inferred compositions are supported by the presence of accessory andesitic, dacitic, and rhyolitic fragments in many of the samples.

Silicic ash-flow tuff - Ash-flow sheets in both the eastern and western facies are rhyolitic, based on phenocryst mineralogy and whole-rock chemical analyses. Phenocrysts are primarily oligoclase, sanidine, and quartz with only traces of ferro-magnesian minerals (Table 1). Chemical analyses indicate generally silicic compositions (Tables 2 and 3), however, the tuffs show very wide compositional variations. These perplexing chemical variations within individual ash-flow sheets have been noted by previous workers (Hay, 1963; Swanson and Robinson, 1968).

Table 1
Modal Analyses of John Day Rocks

12 345
Groundmass92.7a97.6b98.4 94.984.4
Rock Fragments4.00.70.4 -3.6
Sanidinetr-- 1.110.1
Quartztr0.3tr 3.21.6
Plagioclase2.31.41.0 0.8-
(An - Content)35-403510-20 20-25-
Clinopyroxene--tr --
Hornblende--tr tr-
Biotitetr-- --
Opaques0.2tr0.1 tr0.2
Zircontrtrtr -tr
Apatitetrtrtr trtr
1. Tuffaceous claystone and siltstone from western facies (average of 648-207, -S2-9, and -S2-11).
2. Vitric tuff from western facies (648-S7-6)
3. Picture Gorge Ignimbrite, lower cooling unit (average of 648-458a, -458b, and -458c)
4. Basal ash-flow tuff of member A (Swanson and Robinson, 1968)
5. Basal ash-flow tuff of member G (average of 9 specimens)

     a) includes 7.2 percent pumice fragments larger than 0.5 mm
     b) includes 4.7 percent pumice fragments larger than 0.5 mm

- = not recorded

Numerous chemical analyses of the Picture Gorge Ignimbrite and the basal ash-flow sheet of member H indicate a consistent chemical alteration pattern that is dependent on the porosity and crystallinity of the tuff (Table 2). Very densely welded, devitrified zones with little or no observed primary porosity are interpreted to be the least altered parts of the ash-flow sheets. Within these zones analyses are consistent, alkali elements are subequal to each other, water contents are consistently low, and petrographic characteristics do not suggest dissolution of glass or deposition of secondary minerals. Other zones exhibit increasingly severe effects of alteration. Glassy zones are hydrated as indicated by water contents or volatile losses of 3 to 6 weight percent. Typically, hydrated vitrophyres contain less SiO2, TiO2, and Na2O and more Fe2O3, CaO, and K2O than devitrified, densely welded tuffs. Alumina and MgO have no consistent pattern of loss or gain. With minor exceptions such as Fe2O3, the chemical differences between hydrated vitrophyres and devitrified, densely welded tuffs are similar to those reported for Nevada Test Site welded tuffs (Lipman, 1965).

Table 2
Chemical Composition of Fresh and Altered Welded Tuff

Member E
Member H
Picture Gorge Ignimbrite

123 4567 891011
SiO2 76.3777.5183.23 75.9477.576.9 80.8375.3174.07 76.3982.15
TiO2 0.170.390.25 0.350.24
Al2O3 12.0311.398.74 13.0712.211.61 11.3313.0813.08 12.469.22
Fe2O3 0.922.223.23 1.551.03
MnO --- 0.01tr0.04 0.020.03
MgO 0.070.350.03 0.100.04
CaO 1.000.610.31 0.500.650.18 0.210.681.28 0.590.55
Na2O 2.292,921.43 4.413.13.33 2.963.983.60 3.312.41
K2O 5.094.474.77 3.383.954.28 5.144.03
P2O5 --- 0.140.05












LOI 3.741.021.12 0.866.0+0.42 0.641.004.03 1.040.90
1. 648-86, 648-S1-3, 648-S7-10, and DLP 58-50 (Peck, 1964): Hydrated glass.
2. 648-41, 648-S4-4, PTR-71-4b: Devitrified, moderately welded.
3. 648-S1-4: Devitrified, weakly welded.
4. 648-129A and 648-189: Devitrified, very densely welded
5. DLP 58-39A from Peck (1964): Hydrated glass.
6. 648-311: Devitrified, moderately welded.
7. 648-S4-10 and PTR 71-Sb: Devitrified, weakly welded.
8. #3 Hay (1963): Devitrified, very densely welded, fresh.
9. 62-137 and 62-376 from Fisher (1966a); #2 Hay (1963); 648-455B: Hydrated glass.
10. #4 Hay (1963) and 648-458A: Devitrified, moderately welded.
11. 648-455C, 648-455D and 648-455E: Devitrified, unwelded.

- = not detected.

Devitrified, moderately to weakly welded or unwelded tuffs display different chemical changes that become increasingly severe as the porosity increases (Table 2; unpublished data). Typically, SiO2 increases and TiO2, Al2O3, MgO, CaO, and Na2O all decrease with increasing porosity of the sample. Both iron and potassium are somewhat higher in devitrified, moderately welded tuffs but decrease in more thoroughly silicified unwelded portions of the devitrified ash-flow tuffs.

Chemical changes in devitrified samples can be partially correlated with petrographic features. The loss of many elements is consistent with observed leaching of pumice lapilli and phenocrysts from altered specimens. Iron content is largely control led by secondary iron oxides that stain the rocks brick red or form liesegang banding. Very high potassium samples have minute euhedra of potassium-rich feldspar in the groundmass or in pumice lapilli, and very high silica rocks contain abundant prismatic quartz crystals in the groundmass and pore spaces.

The ash-flow tuffs are believed to have been altered by the same diagenetic event that affected the air-fall tuffs and tuffaceous claystones (Hay 1962a; 1963). The very different alteration of air-fall and ash-flow tuffs is due largely to differences in texture. Densely welded non-porous vitrophyres exhibit chemical alteration typical of leaching by groundwater during hydration, such as that reported by Lipman (1965). The chemical effects are generally small, as might be expected for rocks with no primary porosity and little induced porosity.

Porous glassy material from unwelded basal zones of ash-flow sheets and enclosing tuffaceous claystones and vitric tuffs exhibit severe chemical changes induced by large volumes of groundwater acting on large surface areas. The low porosity and permeability of the vitrophyres precluded significant chemical reaction with the groundwater except possibly for hydration and minor leaching. Where cut by fractures that increased permeability and porosity, the vitrophyres have thin (1-5 cm), zenlitic alteration halos along the cracks (Hay, 1963).

Porous, partially welded to unwelded zones in the ash-flow tuffs must have been devitrified prior to diagenetic alteration; otherwise, secondary minerals similar to those in surrounding claystones and tuffs would be expected. Devitrified tuffs contain alkali feldspar and quartz in the groundmass whereas claystone and vitric tuff contain montmorillonite, celadonite, or clinoptilolite. The difference in secondary mineralogy of the welded tuffs and surrounding claystones is probably not due to differences in initial rock compositions because Hay (1963) reported zeolitic alteration of silicic as well as intermediate composition tuff.

The alteration of upper partially welded to unwelded zones is probably unrelated to vapor phase crystallization in the ash flows. This is inferred because typical vapor phase minerals such as tridymite are not commonly observed and quartz is the most abundant silica mineral. Minor chemical changes may have resulted from vapor phase activity, but these were overwhelmed by subsequent diagenetic alteration.

Assuming that our model of consistent chemical change as a function of sample porosity and crystallinity is correct, we can infer the original magma composition of altered units by comparing rocks of similar petrographic character. For example, a rhyolite composition is inferred for the ash-flow sheet of member K, which has no known fresh material, by comparing its composition to petrographically similar altered material from the welded tuff of member H or from the Picture Gorge Ignimbrite (Table 2).

Chemical analyses of altered samples from other units suggest that all of the ash-flow tuffs in the John Day Formation were originally rhyolite in composition (Table 3). Certainly minor chemical differences between ash-flow sheets must have been initially present; however, these are masked by diagenetic alteration effects. In general, the ash-flow tuffs appear to have been moderately alkali rhyolite with silica contents on the order of 75-76%. These rocks appear to be chemically distinct from most of the vitric tuffs and tuffaceous claystones which make up the bulk of the formation.

Table 3
Chemical Analyses of Last Altered John Day Formation Ash-Flow Tuffs

Member A Member CMember E Member GMember H Member IPicture

Lower TuffUpper Tuff


123 456 789
SiO2 74.6873.9975.17 76.3076.6474.06 75.9474.5175.31
TiO2 0.250.610.20
Al2O3 12.5712.7813.30 12.511.3112.45 13.0712.8013.08
Fe2O3* 2.172.693.41 1.202.662.22
MnO -0.030.21 -0.06tr
CaO 0.651.010.59 0.460.761.09 0.500.880.68
Na2O 3.12.743.64 4.412.443.98
K2O 4.386.496.08 5.14.714.37 4.195.703.95
P2O5 -0.20- 0.02-0.11 0.02-.10










LOI 0.620.713.22 1.101.541.81 0.863.461.00
*Total iron as Fe2O3

Sample Identification

1. 648-1a: partially devitrified, densely welded tuff
2. 648-1b: devitrified welded tuff with leached out pumice lapilli and minute quartz and alkali feldspar crystals in groundmass
3. 648-4a: hydrated, densely welded tuff
4. 648-27 and 648-S1-1: crystallized rhyolite flow with leached phenocrysts and quartz and alkali feldspar in groundmass
5. 648-41: devitrified, densely welded, lithophysal tuff with crystal-lined cavities; no thin section
6. 648-34: glassy, partially hydrated vitrophyre
7. 648-129A and 648-189: devitrified, very densely welded tuff with spherulites or very fine grained quartz and alkali feldspar
8. 648-5, 648-245, 648-S2-12: hydrated moderate to densely welded tuff with perlitic texture
9. Analysis 3, Hay (1963): devitrified, very densely welded tuff

Mafic lavas - Although the rocks are somewhat altered reliable chemical analyses have been obtained for most of the mafic lava flows in the formation (Table 4). Member B of the western facies is a trachyandesite with moderate SiO2 content. Equivalent rocks are not known in the eastern facies out crops. Basalt flows in the western facies are all low silica, silica-undersaturated alkali-olivine basalt with high titania. Minor basalt flows in the lower part of the eastern facies have similar compositions.

Table 4
Average Chemical Composition of Mafic Lava Flows

Member B Members E and F
Western Facies
Lower Member
Eastern Facies




H2O+1.41 1.181.98
Member B: average of three analyses from Robinson (1969)

Members E and F: average of 11 analyses in western facies from Robinson (1969)

Lower Member: average of 2 analyses, in eastern facies (Robinson, 1969). Outcrops are area below Picture Gorge Ignimbrite; however, corrrelation to western facies members is uncertain

All samples recalculated water free, all iron as Fe2O3 and totaled to 100%

The trachyandesites and alkali olivine basalts probably represent separate magma pulses, unrelated to each other or to the silicic rocks. They form distinct chemical groups and rocks of intermediate composition are not known. In addition, the high FeO/MgO ratios and Fe2O3 contents of the alkali-olivine basalts indicate an evolution along high-Fe tholeiitic trends rather than along high silica lines toward trachyandesite compositions.


Tuffs, lapilli tuffs, and tuffaceous claystones in the western, southern, and eastern facies are texturally similar to rocks that represent accumulation of ash falls on the land surface (Hay, 1962a; 1963; Fisher, 1966b). These rocks become progressively thicker and coarser-grained from east to west (Waters, 1954; Peck, 1964; Robinson, 1975), suggesting that the source volcanoes lay to the west of the present outcrops. The source volcanoes were largely andesitic to dacitic in composition based on pyrogenic mineralogy. No andesitic or dacitic volcanoes of John Day age are known in the John Day basin, east of the present day Cascades. Previously suggested vents in the Horse Heaven Mining District (Waters and others, 1951) and at Smith Rock (Williams, 1957) are unconformably overlain by John Day rocks (Swanson and Robinson, 1968; Robinson and Stensland, 1979) and are here considered to be Clarno in age. The absence of andesitic or dacitic vents in the John Day outcrop area, the evidence for a western source, and the similarity of these rocks to younger units suggest that the source volcanoes were located beneath the present day Cascade Range. This suggests that Cascade volcanism began about 36 m.y. ago, and that the earliest phases of volcanism were compositionally similar to later ones.

The ash-flow tuffs of the John Day Formation are all rhyolitic in composition, distinctly different from most of the air-fall tuffs and tuffaceous claystones. Most of the ash-flow tuffs occur in the western facies and lateral variations in thickness, intensity of welding, grain size, and abundance of lithophysae all suggest that they were derived from vents west or southwest of their present outcrops. Rhyolitic lava flows and ash-flow tuffs are absent in the western Cascades but numerous large rhyolitic domes occur in the area between the Cascade crest and the Blue Mountain uplift (Peck, 1964; Robinson, 1975; Swanson, 1969). Many of these, such as Powell Butte, Juniper Butte, and unnamed domes near Ashwood and in the Mutton Mountains are known to be John Day in age (Robinson and McKee, unpublished data); others are inferred to be on the basis of lithologic and compositional similarity. Although the ash-flow tuffs cannot be traced to specific vents, the domes and tuffs are compositionally similar and they probably represent the same phase of volcanism. Hence, most of the ash-flows are believed to have been erupted from vents along the eastern margin of the present day Cascade Range and to have spread east and northeast into the John Day Basin. They are largely restricted to the western facies because the Blue Mountain uplift formed a topographic barrier between the eastern and western facies through most of John Day time (Robinson, 1975). The Picture Gorge Ignimbrite in the eastern facies cannot be correlated with any of the ash-flow sheets in the western facies and is believed to have been erupted from an unknown vent in the Ochoco Mountains (Fisher, 1966a). These silicic volcanoes probably also account for the sparse rhyolitic air-fall tuffs and tuffaceous claystones in the lower part of the formation.

The alkali-olivine basalt and trachyandesite flows in the lower part of the formation are most abundant in the western facies in the Antelope-Ashwood-Willowdale area, but they also occur east of the Blue Mountain uplift. These are mostly small, areally restricted flows, clearly of local origin. Scattered dikes and small accumulations of cinders probably mark vents from which some of these flows were erupted (Swanson, 1969; Robinson, 1975),


The John Day Formation consists of three major lithologies, each derived from a different source. Andesitic to dacitic pyroclastic debris in the air-fall tuffs and tuffaceous claystones was derived from volcanoes buried beneath the present day Cascades. Interbedded rhyolitic ash-flow tuffs, lava flows, and domes were erupted from vents lying just east of the Cascades and alkali-olivine basalt and trachyandesite flows were erupted from local vents throughout the John Day basin.

Our model for the evolution of volcanism in northern Oregon is as follows: During the Eocene, andesitic volcanism of Clarno lithology was widespread in north-central Oregon, probably related to vents along the northeast-trending Blue Mountain uplift (Swanson, 1969). In the general area of the present day Cascades, marine sediments were being deposited along with minor basaltic rocks (Peck and others, 1964). Sometime between 40 and 36 m.y. mgo, Clarno volcanism decreased in intensity, probably ending entirely by 36 m.y. Although the age of the uppermost Clarno rocks is not known exactly, no calc-alkaline andesitic rocks of Clarno lithology are known in the John Day Formation.

Initiation of John Day volcanism about 36 m.y. ago signified the emergence of a new volcano-tectonic regime in northern Oregon. Volcanoes in the vicinity of the present day Cascades began erupting andesitic to dacitic pyroclastic material that was deposited in the John Day basin as ash falls. This material represents the earliest documented volcanic activity along the Cascade trend. Simultaneously, vents east of the High Cascades, between the Cascade Range and the Blue Mountain uplift, were erupting rhyolite ash-flow tuffs, lava flows, and minor air-fall tuffs. Most of the ash-flow tuffs were derived from vents west of the present outcrops, whereas lava flows and domes represent local eruptions within the John Day basin. Other vents, mostly east of the major rhyolitic volcanoes, erupted alkaline basalt and trachyandesite.

This pattern of volcanism continued until approximately 25 m.y. ago, at which time the rhyolitic, basaltic, and trachyandesitic eruptions ceased. Andesitic to dacitic eruptions continued in the Cascade Range leading to deposition of the upper part of the John Day Formation.

Deposition of the John Day Formation ceased about 18-20 m.y. ago (Woodburne and Robinson, 1977) coincident with a probable hiatus in Cascade volcanism (McBirney and others, 1974). Following a short period of folding and erosion, the Columbia River Basalt was erupted from vents in the eastern part of the John Day basin beginning about 16 m.y. ago (Watkins and Baksi, 1974; McKee, Swanson, and Wright, 1977). Renewed volcanism in the Cascade Range is reflected in tuffaceous interbeds in the Columbia River Basalt and in post-Columbia River deposits such as the Mascall and Madras Formations.

The regional significance of the pattern of Oligocene and early Miocene volcanism outlined above is not clear. The different magmas may reflect lateral variation in depth of magma generation, in crustal thickness or composition, or in tectonic environment.

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

Last Updated: 28-Mar-2006