Geological Survey Professional Paper 754—C The Portage Lake Volcanics (Middle Keweenawan) on Isle Royale, Michigan

GENERAL DESCRIPTION

The Portage Lake Volcanics on Isle Royale, as on the Keweenaw Peninsula, consists of a thick sequence of basalt and basaltic andesite lava flows with interbedded conglomerate, sandstone, and pyroclastic rocks. Exposures on the island indicate a minimum thickness of 10,000 feet, and as the base of the sequence is not exposed, the total thickness must be considerably greater. The formation is conformably overlain by the Copper Harbor Conglomerate (Wolff and Huber, 1973). The entire sequence is tilted toward the southeast at dips ranging from 55° to less than 5°, with attitudes generally steeper on the north side of the island than on the south. Geophysical data suggest that a major high-angle thrust fault, the Isle Royale fault, lies between Isle Royale and the north shore of Lake Superior (Halls and West, 1971), thus increasing the structural similarity between Isle Royale and the Keweenaw Peninsula.

Records from holes drilled during exploration for copper suggest that there may be more than 150 flows in the section on the island (Lane, 1898), but some units indicated as individual flows in Lane's section may actually be parts of composite flows or duplications caused by faulting. In the same drill logs Lane records 25 clastic units 1 foot or more thick; less than a third of these units are known from outcrop, and those occur primarily in the uppermost part of the section.

VOLCANIC ROCKS

CLASSIFICATION AND NOMENCLATURE

The classification adopted here for the volcanic rocks of the Portage Lake Volcanics on Isle Royale is physically descriptive so as to be useful not only to the specialist but to the interested visitor to the National Park; it is based chiefly upon rock textures. Such a classification has not been widely used outside the Lake Superior region, but by 1898, when Lane's Isle Royale report was published, a textural classification was well established in the Michigan copper district and was used by Lane, who later reviewed the history of the development of this classification (1911, p. 51-66). As emphasized by Butler and Burbank (1929, p. 23-24),

the one outstanding characteristic of the flows that is most useful in their classification is their texture, which serves as one of the major bases of correlation from section to section. Textural classification may be made in hand specimen, drill core, or outcrop. There may be some differences of opinion as to the appropriateness of the names used in this classification but there can be no doubt as to the usefulness of the distinctions.

The following textural classification, based upon Lane (1898, 1911), is used in this report:

Ophite.—Rock with an ophitic texture is produced by crystals of pyroxene that surround and enclose unoriented plagioclase laths, imparting a mottling (fig. 4A). Weathering commonly accentuates the texture, both by increasing color contrast between the pyroxene crystals and the rock matrix and by producing a knobby surface (fig. 5). In freshly broken rock the texture can be distinguished by the flashing of pyroxene cleavage faces. The size of the pyroxene crystals varies with their distance from flow surfaces, becoming progressively larger toward flow interiors; crystals exceed 2 centimeters in diameter in the thick est ophitic flows on Isle Royale.

FIGURE 4.—Characteristic textures of volcanic rocks on Isle Royale. A, Typical ophite. Beach pebble from Mott Island. B, Fine-grained porphyrite, Scoville Point Flow from south side of North Government Island. C, Fine-grained porphyrite tending toward glomeroporphyrite. Tobin Harbor Flow from south side of Porter Island. D, Coarse porphyrite. Huginnin Flow from shoreline just west of Huginnin Cove. E, Pegmatite. From the differentiated zone of the Greenstone Flow on the east end of Passage Island. F, Trap. Minong Flow from the west end of Isle Royale. A, B, D, and E show weathered surfaces.

FIGURE 5.—Coarse ophite with knobby weathered surface. South side of Raspberry Island. Penknife, 7 cm long.

Porphyrite.—Rock with a porphyritic texture is produced by well-defined plagioclase crystals scattered through a finer grained groundmass. The term is applied to two distinct varieties of such porphyritic rocks. One variety has small, blocky, millimeter-sized plagioclase crystals, more or less uniformly distributed through the groundmass (fig. 4B); where the plagioclase crystals tend to clot together, the term "glomeroporphyrite" has been used (fig. 4C). The other variety has larger, tabular crystals more randomly distributed and commonly occurring in clots (fig. 4D); the large crystals are often as long as 2 cm.

Pegmatite.—Rock with a pegmatitic texture is that in which the minerals, especially the plagioclase, are unusually coarse when compared with the other volcanic rock types (fig. 4E). This term, as applied to Keweenawan lava flows, is of relatively recent usage (Cornwall, 1951a); Lane used the term "dolerite," one that is confusing because it is widely used in a different sense.

Trap.—Fine-grained massive rock called trap shows none of the distinctive textures indicated previously (fig. 4F). Lane tended to use "melaphyre" as a rock type and trap for a specific lava flow of that rock type, as for example, the Minong Trap. The term "trap" has also been used in the Michigan copper district to refer to the massive interior part of a flow, as distinguished from the amygdaloidal margins (amygdaloid). In this report "trap" will be used with a textural connotation and "is applied to any nonporphyritic aphanitic rock which is very dark gray or black, regardless of composition. It is thus equivalent to the term 'felsite' as used for lighter colored nonporphyritic aphanitic rocks" (Jackson, 1970, p. 300-301).

The main advantages of a textural classification over a compositional classification for rocks of the Portage Lake Volcanics arise not only from the ease of application in the field but from the nature of the volcanic rocks themselves. Individual flows within the volcanic sequence generally have a characteristic texture throughout; consequently certain stratigraphic units representing flows or groups of flows can be identified and traced in the field on the basis of their texture and thus are excellent stratigraphic marker units. These features are especially important in areas of limited outcrop, as on the west end of Isle Royale or in areas that have been structurally disturbed. On Isle Royale, 12 such flow units have been distinguished within the Portage Lake Volcanics, and these key units provide stratigraphic and structural control for geologic mapping on the island (fig. 6).

FIGURE 6.—Nomenclature for named volcanic flows within the Portage Lake Volcanics on Isle Royale. Shaded areas represent the undivided remainder of the consisting predominantly of ophitic volcanic rocks with lesser amounts of sedimentary and pyroclastic rocks.

COMPOSITION

It has been long recognized that there is a general relation between the texture and the composition of the volcanic rocks in the Portage Lake Volcanics on the Keweenaw Peninsula, although there is considerable overlap. Butler and Burbank (1929) noted that ophitic texture is best developed in the more mafic lavas and that porphyritic and glomeroporphyritic textures are characteristic of more felsic flows or those approaching andesite. Cornwall (1951a) stated that "the ophites are predominantly basalts, whereas the porphyrites, glomeroporphyrites, and melaphyres are predominantly basaltic andesites."

The plagioclase in the ophites is generally labradorite; that in the porphyrites is labradorite or calcic andesine. The pyroxene is predominantly augite with small amounts of pigeonite in some ophitic flows. Ilmenite and magnetite are common accessory minerals. The major secondary minerals occurring both as alteration products and filling original voids include chlorite, sericite, prehnite, pumpellyite, chalcedony, laumontite, and calcite in addition to the opaque alteration products of the olivine.

Primary textures of the volcanic rocks on Isle Royale are well preserved; there is no development of schistosity. Nevertheless, the rocks have been pervasively altered, for of the original minerals, only the pyroxene remains generally fresh. The degree of alteration appears to be much greater on Isle Royale than in similar rocks on the Keweenaw Peninsula, where commonly plagioclase and occasionally olivine are fresh enough for their original compositions to be determined.

In the absence of quantitative mineralogical details from Isle Royale, chemical analyses of representative volcanic rocks are here used to characterize their composition (table 1). The extent to which alteration has affected the bulk chemistry of the rocks is unknown, but alteration is reflected to a certain extent by the presence of appreciable water, probably occurring chiefly in secondary minerals, and by the high oxidation state of the iron. To partially compensate for oxidation, the ferrous-ferric ratio was adjusted to 2:1 prior to calculating the normative minerals in table 1.

TABLE 1.—Analyses of volcanic rocks of the Portage Lake Volcanics, Isle Royale

(click on image for a PDF version)

According to the chemical classification of Irvine and Baragar (1971), which is based in part on normative mineralogy, the suite of rocks from Isle Royale is subalkaline, has tholeiitic affinities (fig. 7), and is predominantly basalt (fig. 8). The apparent variance of their classification with Cornwall's nomenclature results from his use of the classification system of Johannsen (1939), in which the dividing line between basalt and andesite is based solely on the modal anorthite content of the plagioclase (at 50 percent) and under which some of the basalts of this report would be called andesites.

FIGURE 7.—Tholeiitic affinities of analyzed specimens of rocks from the Portage Lake Volcanics on Isle Royale. Classification used is that of Irvine and Baragar (1971). Numbers are keyed to analyses in table 1.

FIGURE 8.—Predominantly basaltic nature of analyzed specimens of rocks from the Portage Lake Volcanics on Isle Royale. Classification used is that of Irvine and Baragar (1971). Numbers are keyed to analyses in table 1. (click on image for a PDF version)

STRATIGRAPHIC UNITS

The topography developed upon the rocks of the Portage Lake Volcanics on Isle Royale consists of a series of ridges and valleys alined in a northeast-southwest direction parallel to the regional strike of the formation (fig. 9). The southeast-facing slopes, controlled by the structural dip, are gentle compared with the northwest-facing antidip slopes and together provide a marked asymmetry to the ridge-and-valley cross profile. Because the sedimentary rocks in the series and the amygdaloidal zones at the base and top of flows are less resistant than the massive flow interiors, the interior parts of the thicker flows are the dominant ridge forming units. The distinguishing texture of an individual flow is generally better developed and more readily apparent in the nonamygdaloidal or less-amygdaloidal flow interior. Fortunately, the part of a flow most useful for identification purposes is the part most likely to crop out. Contacts between flows, on the other hand, are seldom exposed, as they are generally concealed beneath alluvial materials or talus in the depressions between ridges.

FIGURE 9.—Shaded relief map of Isle Royale National Park. Prepared by Alexius J. Burgess (U.S. Geol. Survey, 1957). (click on image for a PDF version)

Because the volcanic rocks on Isle Royale are predominantly ophitic, the nonophitic or the few unusually thick coarsely ophitic flows are most useful as stratigraphic markers. These include finer grained porphyrites and traps which, in addition to their unique textural identity, are generally more resistant to erosion than the ophites and tend to stand out topographically, adding to their usefulness in geologic mapping.

The 12 stratigraphic units within the Portage Lake Volcanics on Isle Royale that have been individually named include five porphyrites, four traps, and three ophites. All names are new except for the Huginnin and Minong Flows, formerly called Huginnin Porphyrite and Minong Trap, and the Greenstone Flow, originally named on the Keweenaw Peninsula. Six of these units can be traced the full length of the island, and two others about two-thirds the length.

The thickness of the units is difficult to estimate, in part because contacts are so sparsely exposed. Some thickness data can be obtained from Lane's (1898) descriptive log of a series of holes drilled across the island near its west end in 1891 and 1892, although correlation problems introduce some uncertainties. Several of the stratigraphic units used in this report appear to correlate with groups of flows in Lane's log rather than with individual flows. It is possible that Lane, in basing his conclusions on the presence of amygdaloidal rock in the core samples, overestimated the number of individual flow tops present or that there is duplication by faulting. Lane was aware of the problem, noting in numerous places in the log that the limits of some flows were very uncertain; in one place he noted that the samples from a core box had been mixed up. It is equally possible that some of the units reported here do indeed consist of multiple flows that are not individually discernible in the field. Each unit has its distinctive rock type, however, and must represent a single volcanic episode that produced rock differing in some respect from that immediately underlying or overlying the unit. The units thus are rock stratigraphic, and all may not necessarily be single flows.

In the discussion that follows, the stratigraphic units are grouped by rock type and not in any specific stratigraphic sequence. In order to provide a stratigraphic framework within which to view the individual units, the sequence is shown in a schematic columnar section (fig. 10). The distribution of some of the units is shown on the bedrock map (fig. 11).

FIGURE 10.—Schematic columnar section of the Portage Lake Volcanics on Isle Royale. Not drawn to scale. (click on image for a PDF version)

FIGURE 11.—Simplified bedrock geology of Isle Royale. Scale too small to show all stratigraphic units described in text. For additional details, see larger scale geologic map (1:62,500) published separately (Huber, 1973b). (click on image for a PDF version)

PORPHYRITE FLOWS

SCOVILLE POINT, MIDDLE POINT, AND TOBIN HARBOR FLOWS

The rock in the Scoville Point, Middle Point, and Tobin Harbor Flows is similar and somewhat unusual; it is not known to crop out elsewhere in the stratigraphic section on Isle Royale. Lane (1898, p. 170), recognizing their distinctive character, informally referred to them together as "the Tobin porphyrites" for occurrences in the Tobin Harbor area.

The Scoville Point Flow is well exposed on the end of Scoville Point, its type locality, at the east end of Rock Harbor. It also is exposed on Edwards and North Government Islands, along part of the north shore of Rock Harbor, along the north shore of Lake Richie, and at scattered localities over the length of the island. On the western part of the island, Red Oak Ridge, which there rivals Greenstone Ridge in height, is preserved because it is capped and armored by the resistant Scoville Point Flow. The rock characteristically has fine equant millimeter-sized plagioclase crystals distributed uniformly through a fine-grained matrix; the texture can best be seen on weathered surfaces (fig. 4B). The unit is at least 100 feet thick at Scoville Point and locally may be as much as 200 feet thick.

The Middle Point Flow is nowhere well exposed. The most accessible outcrop is a small one, considered its type locality, on the south shore of Grace Harbor about half a mile northeast of Middle Point. The flow can be traced northeastward through scattered outcrops for a distance of 12 miles; it appears to be absent for about 9 miles further along strike, and then is exposed over an extent of 2 miles north of Siskiwit Lake. It may not occur on the eastern third of the island. The texture of the Middle Point Flow is similar to that of the Scoville Point Flow except that the plagioclase crystals are less uniform in size and distribution and have a tendency to clot together as in a glomeroporphyrite (fig. 4C). The thickness of the flow is probably somewhat more than 50 feet.

The Tobin Harbor Flow is well exposed on the south arm of Porter Island, along the north side of Tobin Harbor, its type locality, and westward through scattered outcrops for the length of the island. Good outcrops occur on the Mount Franklin Trail just north of Tobin Creek. North of Siskiwit Lake and south of Lake Desor, the Tobin Harbor Flow makes up much of the south slope of Greenstone Ridge. The rock is identical in appearance with the Middle Point Flow, tending toward a glomeroporphyrite (fig. 4C). The flow is probably at least 50 feet thick in the Tobin Harbor area and may be somewhat more than 100 feet thick in the central part of the island.

HUGINNIN AND GRACE ISLAND FLOWS

The rock of the Huginnin and Grace Island Flows is the most distinctive volcanic rock on the island, with its large euhedral tabular plagioclase crystals floating randomly in a fine-grained groundmass (fig. 12). Even as beach pebbles or cobbles, this rock is easily recognized (fig. 13). In referring to the Huginnin Flow, Lane (1898, p. 89) stated that there was no other flow like it in the whole series; except for the Grace Island Flow, which he apparently overlooked, this appears to be true.

FIGURE 12—Porphyrite from Huginnin Flow. Loose slab on beach at Huginnin Cove. Pencil, approximately 15 cm long.

FIGURE 13.—Beach cobble of porphyrite from Grace Island Flow with a rather unusual distribution of plagioclase phenocrysts. Dark spots are amygdules of pumpellyite. From beach on Grace Island; long dimension, 15 cm.

The Huginnin Flow was originally named the Huginnin Porphyrite by Lane for outcrops that occur in the bed of Huginnin Creek just a short distance from where the creek flows into Huginnin Cove near the northwest corner of the island. Although the outcrops in the creek are rather poor, excellent wave-washed outcrops occur on the shoreline about a quarter of a mile southwest of Huginnin Cove. At that locality, the upper contact of the flow is exposed and the plagioclase phenocrysts can be seen to extend undiminished in abundance and size through the amygdaloidal zone to the top of the flow. On the western third of the island, the Huginnin Flow occurs beneath a rather thick cliff-forming ophitic flow and as a consequence is generally buried under talus. However, it can be traced by scattered outcrops from just west of Huginnin Cove to Little Todd Harbor. From Todd Harbor eastward it can be traced for the length of the island, the most accessible outcrops being at the east end of Todd Harbor, at the mouth of McCargoe Cove, and on a small peninsula on the west side of Lane Cove. Lane's drill log gives a thickness of 67 feet for the Huginnin Porphyrite at the northwest corner of the island.

The rock of the Grace Island Flow is indistinguishable from that of the Huginnin Flow. The best and most easily accessible exposures occur on the offshore islands at the west end of Isle Royale proper; on a point south of the campground on Grace Island, its type locality, on Booth Island, and on both ends of Washington Island. On the west end of Washington Island, the base of the flow is exposed, and the plagioclase phenocrysts can be seen to extend through the basal amygdaloidal zone to the bottom of the flow. Outcrops of the Grace Island Flow are not only poor but scarce on the main island, and the unit can be traced by means of very scattered outcrops for about 6 miles eastward from Grace Harbor. That the unit is more extensive is suggested by a single recognized occurrence of the distinctive porphyrite in the same stratigraphic position just beneath the Greenstone Flow on Greenstone Ridge in the central part of the island just north of the Greenstone Ridge Trail about 1 mile east of the east end of Hatchet Lake. On Washington Island and for some distance eastward, the Greenstone Flow is relatively thin, but at about the point where the Grace Island Flow becomes difficult to trace, the Greenstone Flow thickens eastward and becomes more and more a cliff former, burying its own base, as well as the immediately underlying section, beneath massive piles of talus. The Grace Island Flow is probably about 50 feet thick on Washington Island; elsewhere its thickness is unknown.

TRAP FLOWS

EDWARDS ISLAND AND LONG ISLAND FLOWS

The Edwards Island and Long Island Flows, chemically the most basaltic of the rocks analyzed (fig. 8), have important secondary attributes that help in field recognition. With the exception of the upper "chill" zone of the Greenstone Flow, they are the only flows in the stratigraphic section to exhibit well-developed columnar jointing. Lane noted what he referred to as "basaltic" jointing and conchoidal fracture and referred to the rock as "columnar trap." Where the columnar jointing is not well developed, the rock usually exhibits polygonal rather than rectangular jointing, and this, together with fine grain size and conchoidal fracture, aids in field identification.

The type locality for the Edwards Island Flow is the good exposure on the north side of Edwards Island. It is also exposed on adjacent islands and on the point just north of Scoville Point on the south side of Tobin Harbor; columnar jointing is also well developed at these localities (fig. 14). From the Tobin Harbor area, the unit can be traced westward through scattered outcrops for nearly 30 miles to just northwest of Hay Bay. West of Hay Bay the amount of surficial cover on the island increases markedly, and although the Edwards Island Flow may continue farther west, it is concealed. Other accessible areas where the columnar jointing can be observed are the place where the trail from Rock Harbor to Mount Franklin crosses the flow and the south slope of Ransom Hill just west of the trail from Daisy Farm Campground to Mount Ojibway. In thickness the Edwards Island Flow probably averages about 50 feet.

FIGURE 14.—Columnar jointing in Edwards Island Flow on north side of Edwards Island.

The Long Island Flow is exposed on a chain of small narrow islands in Tobin Harbor, including Long Island, its type locality. Columnar jointing is not noticeable in the Tobin Harbor area, but polygonal jointing is well developed on Long and Third Islands. Although the Long Island Flow can be traced as far west as the Edwards Island Flow, to just northwest of Hay Bay, outcrops are generally very poor, and at only one relatively inaccessible locality was well-developed columnar jointing observed. About 1-1/4 miles west of Lake Richie along the north fork of the creek draining into the west end of the lake, columns form a cliff about 25 feet high. The Long Island Flow commonly contains sparse small bluish agate amygdules that help distinguish isolated outcrops of this unit from those of the Edwards Island Flow, in which agates have not been observed. Because the islands where the flow is best exposed project only a few feet above water level, a stratigraphic thickness of no more than 25 feet is visible; however, the flow is probably less than 50 feet thick.

MINONG FLOW

The Minong Flow was originally named the "Minong Trap" by Lane (1898, p. 16) for its occurrence at the Minong mine. The rock is dark and fine grained and has a pronounced conchoidal fracture (fig. 4F). Lane noted a tendency toward "basaltic" (columnar) jointing, but no well-developed columns were observed in the present study. Because it is thick and resistant, the Minong Flow is an excellent ridge former and can easily be traced the length of the island. In fact, Minong Ridge is generally characterized by the precipitous antidip slopes on its north side. The rock is further characterized by the presence of sparse bluish-white agate amygdules. The Minong Ridge Fire Manway (trail) follows the Minong Flow for much of its length and thus makes numerous outcrops of the flow readily accessible. Excellent wave-washed outcrops are found on some of the small promontories jutting into Lake Superior, especially north of Locke Point, in Five Finger Bay, and at the far west end of Isle Royale. At the west-end locality, the base of the unit is exposed and the rock is uniformly dense down to the contact, with almost no amygdaloidal zone present. The thickness of the Minong Flow is 77 feet in Lane's drill log.

Lane (1898, p. 82) also described and named a "Minong Porphyrite," which he designated more felsic than the rock he described as "Tobin porphyrite" and which occurs immediately above his "Minong Trap." In fact, he considered the porphyrite and trap possibly to be magmatic differentiates of a single flow. In describing the porphyrite from drill samples, Lane noted "the extreme fineness of grain, the scoriaceous, porous and brecciated appearance, peculiar in that the pores are fine and irregular." Lane's Minong Porphyrite, now believed to be of pyroclastic origin, is discussed in a later section.

AMYGDALOID ISLAND FLOW

The Amygdaloid Island Flow, an andesite, is the most felsic volcanic rock of those analyzed (fig. 8; table 1). It is exposed only on Amygdaloid Island near the base of the exposed stratigraphic section of the Portage Lake Volcanics. Like the other traps, it is a resistant rock and forms the highest ridge running the length of Amygdaloid Island. Some of the best and most accessible exposures are those on the promontory south of Crystal Cove at the east end of the island, its type locality, and at the natural bridge in the central part of the island—a wave-cut arch formed at a time when a higher lake level existed in the Superior basin (Huber, 1973a). One of the most distinctive aspects of the rock is the presence of rather abundant agates with a characteristic flesh-pink cast and commonly with centers of massive or drusy quartz (fig. 15). The Amygdaloid Island Flow is probably less than 50 feet thick.

FIGURE 15.—Specimen of Amygdaloid Island Flow with characteristic agate amygdules (sawed surface). From northeast end of Amygdaloid Island.

OPHITE FLOWS

The volcanic rocks in the Portage Lake Volcanics on Isle Royale are predominantly ophitic. The stratigraphic marker units described to this point are non ophitic and thus are useful because of their conspicuous contrast within the generally ophitic sequence. Three ophitic flows on Isle Royale, however, are useful marker horizons because they also stand out from the mass of ophitic flows: the Greenstone, Hill Point, and Washington Island Flows.

GREENSTONE FLOW

Lane (1898, p. 98) referred to the Greenstone Flow as "the Greenstone, the 'backbone' [of the island] and biggest ophite of all." Application of this formal name to the flow on Isle Royale resulted from Lane's correlation of this flow with the type Greenstone of the Keweenaw Peninsula, the thickest ophitic flow there. The name "Greenstone" eventually evolved into "Greenstone Flow" on the peninsula, and as the correlation with Isle Royale is still considered valid, the name Greenstone Flow is also used on the island. The Greenstone Flow on Isle Royale, as on the peninsula, has undergone magmatic differentiation, and this feature primarily, in addition to thickness, led to the correlation. Extensive descriptions and analytical data for the Greenstone Flow on the Keweenaw Peninsula were presented by Broderick (1935) and Cornwall (1951a, b).

As late as 1911, some workers considered the "Greenstone," both on the Keweenaw Peninsula and on Isle Royale, to be an intrusive sill because of its thickness and differentiated character (Van Hise and Leith, 1911, p. 381, 389). Lane (1911, p.939-940), however, argued convincingly for its origin as a lava flow, an interpretation confirmed by subsequent work on the peninsula based on thousands of feet of diamond-drill samples (for example, Broderick and others, 1946, p. 683-684).

The Greenstone Flow is indeed the backbone of Isle Royale, for it forms the most prominent ridge running the length of the island, Greenstone Ridge. The Greenstone Ridge Trail, which follows the ridge for nearly its entire length, provides access to numerous scattered outcrops, but only at the far east end of the island near Blake Point is a reasonably complete cross section of the flow exposed. There the flow can be seen to consist of four units with the following approximate thicknesses: A lower ophitic zone (100 ft), a central pegmatitic zone (75 ft), an upper ophitic zone (175 ft), and an uppermost columnar-jointed trap (50 ft), for a total thickness of about 400 feet. The Greenstone Flow is thickest in the central part of Isle Royale, where it is estimated to be nearly 800 feet thick. It thins westward to less than 100 feet on Washington Island.

The lower ophitic zone, which makes up the cliffs of The Palisades near Blake Point at the east end of the island, shows an increase in grain size of the ophitic augite crystals from a few millimeters near the base (base not exposed) to as much as 1-1/2 cm in the upper part of the zone. The contact between the lower ophitic zone and the pegmatitic zone is not exposed but is considered to be fairly abrupt, as it can be projected with good continuity between outcrops of the two rock types for a distance of more than 30 miles and at places can be located within a stratigraphic interval of 25 feet. The pegmatitic zone is extremely variable in texture and grain size but characteristically has an interlocking mat of elongate plagioclase crystals, which are as much as 2 cm long (fig. 4E). The contact of the pegmatitic zone with the upper ophitic zone is gradational over a few tens of feet, with irregular patches of the two rock types intermingling. The upper ophitic zone is fairly coarse at its base, with augite crystals as much as 1 cm in diameter, but grades upward rather rapidly into finer ophite.

The columnar-jointed trap is exposed on the chain of small islands south of Merritt Lane, on Red Rock Point, and at a few additional isolated localities on the main island. The trap was not seen in actual contact with the upper ophite, for the contact between them apparently is a zone of low resistance to erosion and everywhere occupies a physiographic depression. The trap is considered to be part of the Greenstone Flow, rather than a separate overlying flow, by analogy with the type Greenstone Flow on the Keweenaw Peninsula, where a columnar-jointed "melaphyre" is typically found near the top of the flow (Cornwall, 1951a; 1954a, b, c). Compositionally the trap on Isle Royale plots closer to the ophite of the Greenstone Flow than to other columnar-jointed trap in the section (fig. 8) on the basis of one analysis of each that is available.

The lower ophite forms the commonly precipitous, antidip slopes on the north side of Greenstone Ridge, whereas resistant pegmatite commonly armors the gentler, south dip slope. The contact between the lower ophite and the pegmatite usually is near the crest of the ridge; because it is a relatively sharp contact, it provides a very useful internal marker horizon, especially for the recognition of offset along faults that cross the ridge. The pegmatitic zone can be traced in outcrop from Blake Point as far west as Mount Desor, west of which outcrops are sparse. Six miles west of Mount Desor, where the flow is perhaps 265 feet thick, the pegmatite appears to be poorly developed, according to Lane's drill log. At the west end of Isle Royale and on Washington Island, the pegmatitic zone appears to be absent. The pegmatitic zone is likewise missing on the Keweenaw Peninsula, at places where the Greenstone Flow is thin (Davidson and others, 1955).

HILL POINT FLOW

The Hill Point Flow is a typical ophite useful as a stratigraphic marker only because of its thickness and exceptionally coarse grain size; it contains poikilitic augite crystals as much as 2 cm in diameter. The flow is 158 feet thick near the west end of the island, according to Lane's drill log. It crops out in imposing cliffs along the north shore of the island from Huginnin Cove to McCargoe Cove and then along the south side of Pickerel Cove eastward to Hill Point, the type locality. A few thin seams of pegmatite can be seen near McCargoe Cove, but they are not extensive enough to help in identifying the flow elsewhere.

WASHINGTON ISLAND FLOW

The rock of the Washington Island Flow is difficult to describe but distinctive when seen; perhaps Lane's application of the term "glomeroporphyritic ophite" suggests the problem. The rock is clearly an ophite, as indicated by the poikilitic augite crystals, but the ophitic texture is somewhat obscured by the fact that the plagioclase crystals are somewhat larger than in the typical ophites and have about the same distribution and size within the augite crystals as in the ground mass. Other distinctive features are occasional larger porphyritic plagioclase crystals, about one-half centimeter in average size, and a pervasive greenish speckled appearance caused by the rather uniform distribution of chlorite (fig. 16). In thin section the rock exhibits a relict diktytaxitic texture, chlorite and other secondary minerals filling what appear to have been interstitial voids, as well as replacing earlier minerals, such as olivine.

FIGURE 16.—Specimen from Washington Island Flow. Ophitic texture is obscure, but characteristic dark chlorite splotches are visible. Compare with figure 4A.

The Washington Island Flow is well exposed at its type locality on the south side of Washington Island, where it caps the highest ridge. On the main island it can be traced eastward from Grace Harbor for about 5 miles, beyond which it appears to pinch out as the Greenstone Flow increases in thickness. A single out crop of similar rock in the same stratigraphic position above the Greenstone Flow occurs near the east end of the main island on a small point on Porter Island that partially separates a narrow bay from an interior cove. The unit is more than 200 feet thick on Washington Island but thins to about 80 feet 5 miles to the east; it can be no more than a few tens of feet thick on Porter Island.

SEDIMENTARY ROCKS

Epiclastic sedimentary rocks crop out only in the upper part of the Portage Lake Volcanics, although they are known from drill records to occur lower in the section. Those that crop out are all above the upper-most of the named volcanic units, the Scoville Point Flow, and are best exposed in the Chippewa Harbor area, where seven units of pebble conglomerate and sandstone are interbedded with a series of generally thin ophitic lava flows (fig. 17). These sedimentary units can be traced eastward to a point where most of them disappear beneath the waters of Moskey Basin and Rock Harbor; two are exposed discontinuously on the chain of small islands on the south side of Rock Harbor. All the units can be traced westward from Chippewa Harbor as far as Siskiwit Lake, but farther west outcrops are sparse and correlation of individual units is uncertain. The most accessible outcrops are those in the Chippewa Harbor area, on the south side of Siskiwit Lake, the north side of Conglomerate Bay, the south side of Moskey Basin, and the south side of Mott Island. Conglomerate can been seen on the dumps at the Island Mine, northwest of Siskiwit Bay.

FIGURE 17.—Columnar section of upper part of Portage Lake Volcanics showing distribution of sedimentary units in the Chippewa Harbor area. All the ophitic flow zones consist of multiple flows.

The sedimentary units vary in thickness and in grain size along strike, but on the whole the units seem surprisingly uniform in thickness, considering their probable lateral extent, more than 40 miles. For example, the uppermost unit is predominantly pebble conglomerate for its entire length and appears to thicken eastward very slightly. It is about 30 feet thick west of Siskiwit Bay and at Hay Bay and Chippewa Harbor and increases to about 40 feet on Mott Island. The lower-most sedimentary unit, the Island Mine Conglomerate Bed (Lane, 1898, p.99), is a pebble conglomerate about 20 feet thick south of Washington Harbor and at the Island Mine and the west end of Siskiwit Lake; it changes to a sandstone also about 20 feet thick at Chippewa Harbor, and it is a sandstone at Moskey Basin, probably slightly thinner.

The pebbles in the conglomerate are essentially all of volcanic rock, with mafic varieties about twice as abundant as felsic ones. In this respect the conglomerate differs somewhat from the Copper Harbor Conglomerate, higher in the section, where felsic varieties predominate (Wolff and Huber, 1973). The sandstones are very feldspathic, plagioclase being the most abundant among granular constituents. Fine-grained hematite is ubiquitous, giving all the sedimentary rocks a reddish cast and generally obscuring other fine-grained material in the matrix.

The conglomerates and coarser sandstones commonly exhibit a scour-and-fill structure (fig. 18), suggesting an origin in a braided stream system. Locally the sedimentary units are fine-grained even-bedded sandstone throughout and probably represent flood plain deposits. The presence of desiccation cracks and raindrop impressions also point toward a fluvial depositional environment. Paleocurrent data are too scarce to treat statistically, but foreset crossbeds and pebble imbrication suggest that the direction of sediment transport was down the present structural dip, or generally southeast toward the axis of the Lake Superior syncline. This conclusion is compatible with paleocurrent data for the Copper Harbor Conglomerate (Wolff and Huber, 1973).

FIGURE 18.—Scour-and-fill structure in pebbly sandstone near the west end of Chippewa Harbor. Handle of penknife approximately 7 cm long.

A sandstone unit below the Long Island Flow, a conglomerate immediately beneath the Greenstone Flow, and several conglomerates below the Hill Point Flow are indicated in Lane's drill log and described as being similar to those mentioned; none of these are known from outcrop.

Another type of rock, with at least a sedimentary component, is one known in the Michigan copper district as a "scoriaceous conglomerate" or "amygdaloid conglomerate," or even "scoriaceous amygdaloid" (Lane, 1911, p. 69). This type consists of irregular-shaped fragments of highly amygdaloidal rock or scoria in a matrix of fine sand or silt and might more properly be called a breccia (fig. 19). The percentage of fragments diminishes stratigraphically upward. The rock was formed when fine sedimentary material swept over the fragmental top of a lava flow and incorporated in the resulting deposit various-sized fragments from the underlying flow top. Lane described several units of this type in his Isle Royale drill log, but only one is well exposed at the surface. It occurs immediately above the Amygdaloid Island Flow and can best be seen in a mile long exposure along the south shoreline of Amygdaloid Island near the east end of the island.

FIGURE 19.—Breccia occurring above the Amygdaloid Island Flow. Specimen from south shore at east end of Amygdaloid Island.

PYROCLASTIC ROCKS

Lane recognized that some of the so-called sedimentary rocks in the Portage Lake Volcanics on Isle Royale are of pyroclastic origin or contain appreciable pyroclastic debris. He referred to these rocks as ash beds and tuffs and described fragments having "the conchoidal forms of glass ash" (Lane, 1898, p. 171-172). Unfortunately, the pyroclastic rocks, like the epiclastic rocks, rarely crop out and are therefore of little use in the field as stratigraphic markers. A tuff-breccia immediately overlying the Greenstone Flow is the single exception; it is exposed sporadically for the length of the island and indeed helps to identify the Greenstone Flow on Washington Island, where that flow is thin and lacks much of its distinctive character, looking much like any other ophite. Other pyroclastic rocks that occur just above and below the Minong Flow rarely crop out.

The basal part of the tuff-breccia overlying the Greenstone Flow contains much rubble that appears to have been obtained from the Greenstone Flow (fig. 20). Both fragments and matrix within the main body of the tuff are very line grained, suggesting that much of the rock was originally glassy. Relict shard structures and collapsed pumice fragments are sometimes visible in thin section, and many rock fragments are scoriaceous. Other rock fragments exhibit piotaxitic texture. Corroded quartz and plagioclase crystals are common, and staining techniques reveal appreciable K-feldspar in the fine-grained matrix, indicating a rather felsic composition. From the east end of the island to Mount Desor, the tuff is also characterized by numerous agates with a pink or red cast (fig. 21). Unlike the ovoid agates typically occurring in many of the volcanic flows, these agates tend to have a very irregular shape similar to the "thunder eggs" occurring in tuffs of the Columbia River Plateau in the Western United States. Such agates have been interpreted by Ross (1941) as chalcedony-filled spherulitic cavities in welded tuff, and their presence in the tuff on Isle Royale, together with other small ellipsoidal vesicles now filled with secondary minerals, suggests that the pyroclastic material was hot when deposited. The tuff where exposed on Washington Island contains quartz and plagioclase crystals, relict shards and pumice, and scoria fragments, but no agates, amygdules, or other evidence of deuteric thermal activity. Perhaps it was far enough away from the pyroclastic source to have been deposited at a lower temperature than the tuff to the east. The most accessible outcrops of the tuff-breccia are on the north side of Tobin Harbor near Newman Island. Its maximum thickness is probably between 25 and 50 feet.

FIGURE 20.—Tuff-breccia overlying the Greenstone Flow. Fragmental character is clearly evident in this outcrop, which is close to the base of the pyroclastic unit on the north shore of Tobin Harbor near Newman Island. Knife, 7 cm long.

FIGURE 21.—Agate typical of those occurring in the tuff-breccia overlying the Greenstone Flow. Knife, 7 cm long.

The tuff-breccia above the Minong Flow is similar to that above the Greenstone Flow, with abundant spherulitic structures, some containing agate. This unit is probably the same as Lane's Minong Breccia (or Minong Conglomerate) together with his Minong Porphyrite (1898, p. 82, 87). Maximum thickness probably does not exceed 25 feet. The only known exposures of this unit occur in a belt about 1 mile long at the far west end of the island and on the peninsula east of Five Finger Bay at the east end of the island, although it was reported from the workings of the Minong Mine.

Below the Minong Flow a single exposure on a shoreline cliff at the west end of the island shows an ash deposit only a few inches thick in its now-compacted state. It appears to be an air-fall deposit; shard structures are beautifully preserved, and no rock fragments were noted in the few samples examined. The upper part of the ash deposit was later fused by the heat of the overlying Minong Flow, with the downward development of spherulitic structures and the creation of a compaction foliation. A rhyolitic composition is indicated by the refractive index of fused glass beads of the tuff.

STRATIGRAPHIC SUMMARY

The individually named flows of the Portage Lake Volcanics on Isle Royale have been described in some detail, along with the other assorted rock types that occur in the sequence. The stratigraphy of the largely volcanic pile has been summarized in a schematic columnar section presented earlier as figure 10, a composite from various parts of the island; not all units shown are everywhere present. Most of the clastic interbeds are known only from drill core, and such data are limited to the western part of the island. The columnar section is not drawn to scale, but some sense of scale can be obtained from the geologic section (fig. 22).

FIGURE 22.—Geologic section of Isle Royale in the vicinity of Lake Desor and Houghton Ridge. Letter symbols represent units shown in figure 10, with the addition of psc, sandstone and conglomerate; pp, pyroclastic rock; pu, Portage Lake Volcanics, undivided. (click on image for a PDF version)

STRATIGRAPHIC THICKNESS

Some general data have been presented regarding the thickness of individual stratigraphic units, but for most units data on thickness variations along strike are scanty. Because of the hogback asymmetry of ridges formed by individual flows and the tendency of the upper amygdaloidal parts of flows to be eroded, the base rather than the top of most units can be more precisely located even when both contacts are concealed. By utilizing the bases of certain key flows as marker horizons, thickness variations for larger segments of the Portage Lake Volcanics can be calculated for a number of sections across Isle Royale with considerable reliability; the results are presented in a longitudinal stratigraphic section with the base of the Greenstone Flow as a datum plane (fig. 23). Some additional data on the section below the Hill Point Flow, in the vicinity of Amygdaloid Island, have been obtained by tracing on aerial photographs certain topographic lineaments presumed to be stratigraphic horizons.

FIGURE 23.—Longitudinal stratigraphic section showing variations in thickness of the Portage Lake Volcanics on Isle Royale. Data on the Copper Harbor Conglomerate from Wolff and Huber (1973). (click on image for a PDF version)

The longitudinal section brings out two significant features. First, the Greenstone Flow is thickest near the central part of Isle Royale and thins toward both the west and east ends of the island. Secondly, the entire exposed section of the series on Isle Royale also thins toward both ends from a maximum in the center. The Copper Harbor Conglomerate has a similar geometry at the west end of the island. (See Wolff and Huber, 1973, for details.) These features may have some bearing on the topographic expression of Isle Royale as a ridge in the Lake Superior basin, as one would expect the entire section to be more resistant to erosion and stand topographically higher where the individual flows are thickest.