top of page

Report Archeological Survey, Bear Paw by Dr. Douglas Scott


No formal archeological inventories have been conducted at Bear Paw battlefield or in the immediate area. However, the battlefield has been the focus of research and collecting efforts by numerous individuals for many years. Only a few of these collecting efforts are documented. The most extensive collecting efforts were undertaken by the late Thain White of Dayton, Montana, Gordon Pouliot of West Glacier, Montana, and Norman Johnson of Havre, Montana. Their efforts are well documented and constitute a primary data set of the types and quantities of camp and battle debris left on the field after the battle (Scott 1997). Local researchers Leroy "Andy" Anderson of Chinook and Paul English of Havre have also identified a number of rock cairns and rock alignments both within and adjacent to the park boundary, some of which yielded metallic debris when they conducted some preliminary metal detecting in the area after the 1991 range fire.

This report documents a surface reconnaissance level inventory conducted between June 5 and 8, 2000 within the General Management Plan defined boundary. The work was confined to visual surface inventory only.


Project Area Location and Field Methods

The Bear’s Paw Battlefield is situated within the open, moderately rolling prairie country of north central Montana along and adjacent to a portion of Snake Creek. The battlefield site is situated in the foothills of the north slope of the Bear’s Paw Mountains and occupies portions of two distinct topographic features. The area where the Nez Perce were camped when attacked by the Fifth Infantry, and Second and Seventh Cavalries is on a level, primary terrace of Snake Creek. The prairie edge and surface to the east and west of Snake Creek is elevated from 20 to 40 feet above this primary terrace and on the east, is dissected by several ephemeral drainage systems which trend toward the Nez Perce camp locale from the east and southeast. The majority of the battlefield site, and the primary locations where the U.S. Army units took up positions against the Nez Perce are on this prairie surface above the creek bottom.

Field Methodology

The fieldwork consisted of the crew walking the ground in a series of parallel transects (Figure 1) until the entire park area and adjacent lands identified as part of the General Management Plan boundary were covered (Figure 2). Transect spacing varied depending on vegetation density. Spacing was approximately 3 meters in areas with good ground visibility and extended to 10 meters between individual crew members when grass cover was so dense as to obscure the ground surface. Approximately 1400 acres were covered at a reconnaissance level during the fieldwork. Nearly every area within the GMP boundary was covered with a dense vegetative matte. Small portions of the northwestern or Miles cannon pit area and some areas east of the park boundary had ground visibility that approached 50%. Ground visibility was generally less than 20% in most areas. The fieldwork can only be characterized as a reconnaissance level. Thus only the most visible features were seen and recorded.

Dan Foster of NEPE ably guided the crew and maintained interval spacing. The crew was composed of Harold Roeker (MWAC), Dick Harmon (Volunteer), Kelsey Attenhoffer (Volunteer), Connie Constan (Volunteer), Jim Magara (Volunteer), David Thorn (Volunteer) for the entire period of fieldwork. Jon James (BIHO) and Tom James (Volunteer) assisted in the inventory for two days, and the crew was joined by NEPE Superintendent Doug Youri, Otis Halfmoon, Kathleen Halfmoon, Carol High Eagle, and Arthur Currence (BEPA) for one day.

Standard MWAC archaeological data‑recording methods were used in each component of the operation as specified below. Individual artifacts were recorded and noted in the electronic data log. Field notes were also used to record field data. Exposed in‑place artifact specimens and topography were photographed and recorded digitally.


After covering several transects recording was begun. Each artifact or feature marked by a pin flag and piece‑plotted as follows. The instrument was set up at a selected grid coordinate marker determined by a PLGR global positioning device. Distance and azimuth readings for each artifact or feature were recorded in reference to the known grid coordinates. A Sokkia total station transit was used to record the features within the park boundary. The GPS unit was used to record coordinates of features observed outside the park boundary.

Once the data was collected, electronically, it was downloaded each evening into a laptop computer using the Sokkia MAP software and AutoCad Light 98 to develop a field map. That data was edited and AutoCad Light 98 used for final map production for the report.


Prehistoric and Historic Archeology Background

Prehistoric-Protohistoric Cultural Sequences in Northern Montana

The following information on the prehistory of the area is borrowed from the Bear Paw prehistoric archeological overview prepared by Rennie and Brumley (1994). Their overview and assessment of the region's prehistoric archeology is the best available and defines the current knowledge of the area as well as defining some excellent research questions. Archaeologists working on the northwestern Plains, which encompasses the project area, have found evidence of human occupation extending back over at least the last 11,000 years. The reader is referred to Frison (1991), Reeves (1969, 1973, 1983), Ruebelmann (1983), Brumley and Dau (1988), Vickers (1986) and Brumley and Rennie (1993) for a thorough discussion of various models developed for interpreting the region's prehistory.

Each cultural phase or complex within the region is defined largely on the basis of one or more distinctive projectile point styles or types. Ceramics appear intermittently within the prehistoric cultural record of the northwestern Plains including northern Montana during the last 2,000 years; and, where present,  provide a second major criterion for cultural complex/phase definition. Relative and absolute dating of these phases or complexes is based on excavation of sites where such diagnostic forms have been found in stratified deposits. Organic material, such as bone and charcoal, found associated with various cultural assemblages can be radiocarbon dated, providing a close approximation of the actual age of various materials and the relative age of materials from non‑stratified sites. The three major cultural periods, and their general defining characteristics are:

Early Prehistoric Period

Dating from ca. 11,000 to 7,700 radiocarbon years BP, this period contains a number of archaeological units characterized by projectile point styles presumably designed for use on a heavy throwing or stabbing spear. Early Prehistoric period complexes recognized in northern Montana include Clovis, Goshen, Agate Basin, Hell Gap, Alberta, Cody and the Plains/Mountain and/or Foothill Mountain complexes. Early Prehistoric period point types are generally variations on lanceolate or stemmed forms. Early Prehistoric or Paleo Indian peoples appear to have been primarily big game hunters with the earliest well defined groups (Clovis/Goshen) known elsewhere to have hunted a variety of now extinct animal species including mammoth. Later groups generally relied on early forms of bison as their primary food source. Evidence of stone boiling in the form of water fractured, fire cracked rock is presently lacking for the Early Prehistoric period. Evidence of Early Prehistoric period people in northern Montana is limited, and is presently restricted to a number of widely scattered surface finds which are only interpretable by reference to excavated materials from other regions.

Materials similar to that identified elsewhere in the northwestern Plains as Clovis or Goshen Complex materials is represented in northern Montana by an isolated point from northern Blaine County (Deaver 1980)  and points recovered from the surface of four campsites situated within and adjacent to the Bears Paw Mountains in Choteau, Blaine and Hill Counties (Brumley 1988a). In addition to Clovis, an essentially complete sequence of projectile points characteristic of the other Early Prehistoric period complexes have been recovered at several surface sites throughout the Bears Paw Mountains. Brekke (1970) describes surface finds from an extensive campsite located around a major spring in central Blaine County, just south of the Milk River valley. Recovered materials there include Hell Gap points as well as point types characteristic of Middle and Late Prehistoric period cultural complexes. Within the Milk River valley to the east of Fort Belknap, Rossillon (1985) notes finding an isolated Paleo-Indian point she identified as Agate Basin. Ruebelmann (1983) reports on an Agate Basin point and Deaver (1980) on an Alberta point found on the surface at separate site localities within Phillips County to the east of the Little Rockies. No evidence of Early Prehistoric occupation has yet been reported within or in the immediate vicinity of the project area.

Middle Prehistoric Period

Dating from ca. 8,000 - 1,300 BP, this period is characterized by projectile point types presumably designed for use with the spear thrower or atlatl. Major complexes or phases include Mummy Cave, Oxbow, McKean, Pelican Lake, Yonkee, Sandy Creek and Besant. Cultural groups during this period were predominately bison hunters. However, certain contemporary groups occupying the area of central and southern Montana and Wyoming appear to have developed a more diversified subsistence economy based on hunting a broader spectrum of animal species, as well as gathering and processing wild plant foods. Evidence of stone boiling in the form of water fractured, FCR is abundant throughout both the Middle and Late Prehistoric periods. The presence of this evidence appears to reflect a major adaptive change in food processing and storage technologies. Pottery first appears in some Besant phase sites during the latter part of the Middle Prehistoric period.

Early side-notched atlatl points, characteristic of what is identified as the Mummy Cave complex, are well represented in surface finds at a number of site locales within and adjacent to the Bears Paw Mountains (Brumley 1988 a, b).

The McKean complex is primarily characterized by the presence of McKean Lanceolate, Duncan and Hanna atlatl point forms (Wheeler 1952). Another projectile point form referred to as Mallory has been found, on occasion, in McKean complex assemblages from southern Montana to northern Colorado and western Nebraska (Forbis nd; Lobdell 1973; Morris et al. 1984; Munson 1990; Reher 1979). Mallory points are not presently known from northern Montana. From central and southern Montana south, a number of McKean complex sites have produced grinding slabs and rock lined hearths. These artifacts and features are commonly interpreted as reflecting an increased reliance on plant foods during McKean complex times (Frison 1991; Keyser 1986). Further north on the Canadian Plains, similar evidence is lacking with McKean peoples predominately hunting bison for subsistence (Brumley 1975, 1978). In northern Montana, McKean Lanceolate, Duncan, and Hanna projectile points are common in surface collections. However, no excavated or dated cultural assemblages have as yet been reported.  

The Pelican Lake complex is recognized by most researchers as characterized by atlatl size, corner-notched projectile points. A small number of Pelican Lake assemblages contain a few very small corner-notched, arrow point size projectiles, which suggest the limited presence or coexistence of the bow and arrow along with the atlatl. The few  presently available radiocarbon dates for the Pelican Lake assemblages containing these small points are all from sites situated on the Canadian Plains. These dates are quite early in terms of the overall time span for the Pelican Lake complex, suggesting the atlatl was not just being replaced by the bow, but may have coexisted with it for a long period of time as a minor or secondary weaponry system. 

Brumley and Rennie (1993) note a trait dichotomy within the northern and southern expressions of the McKean, Pelican lake and Avonlea phases. Southerly (from ca. central Montana south) assemblages of these complexes/phases are characterized by the occasional to frequent presence of rock lined hearths and/or grinding slabs. These traits are presently, totally absent from more northerly assemblages of these phases/complexes.

Within the general study area, Davis and Stallcop (1965) report on excavations conducted at the Keaster site (24PH401), a multi‑occupational bison kill site located in the south of Phillips County, a short distance west of the Little Rockies. Another poorly known cultural complex referred to as Sandy Creek has been defined by Dyck (1983). Present evidence suggests Sandy Creek to be temporally intermediate and related to both earlier Oxbow and later Besant. Sandy Creek is characterized by rather non-descript, side-notched atlatl points which, out of dated contexts, could fit into the range of variation of points found associated with Mummy Cave, Oxbow and Besant complex assemblages. Such non-descript points are abundant in surface collections from the Bears Paw Mountains (Brumley 1988 a-b), and may reflect the presence of this phase within northern Montana.

Particularly abundant in northern Montana during the latter part of the Middle Prehistoric period are materials referable to the Besant complex. The primary diagnostic of the Besant complex are atlatl sized, shallow side-notched projectile points. The range of variation in point form from some Besant assemblages is broad and appears to overlap in part with projectile points of the earlier Mummy Cave complex and Sandy Creek phases, to which Besant peoples have been suggested to be related (Brumley and Dau 1988; Brumley and Rennie 1993).

Herdegen's Birdtail Butte site (24BL1152) is a small but extensively utilized campsite and bison jump located near the southeast margins of the Bears Paw Mountains. Diagnostic projectile points recovered from the surface in the campsite portion of the site include materials referable to the McKean, Pelican Lake, Besant, Avonlea, and Prairie/Plains complexes of the Middle and Late Prehistoric periods. Test excavations within the bison jump portion of the site revealed a series of 16 stratigraphic units within four meters of sediments.

Small, arrow point sized versions of Besant projectile points have been found in some late Besant assemblages and are referred to as Samantha points. The presence of Samantha points is believed to reflect late Besant assemblages, when the transition from use of the atlatl to the bow and arrow was taking place. Ceramics have also been found in some Besant assemblages. It appears that ceramics become more frequent in Besant assemblages as one proceeds east toward the Dakotas. In northern Montana, the only  Besant ceramic association known to the junior author is from a surface blowout site along Tiber Reservoir (Brumley nd).

Late Prehistoric Period

Dating from ca. 100 AD to Historic times, this Period is characterized by projectile points clearly intended for use with the bow and arrow. Point forms include a variety of un-notched, stemmed and notched forms. Bison hunting remained the primary subsistence activity. Communal bison kills which involved coordinated efforts by groups of hunters driving number of animals over cliffs, into corrals, or into natural traps, are present throughout the entire archaeological record. They appear, however, to reach a peak in both number and magnitude during the Late Prehistoric.

The Avonlea complex marks the introduction of the bow and arrow as the dominant weaponry system on the northwestern Plains. The complex is defined largely on the basis of a form of delicate, very well made, side-notched arrow point. Small, well made corner-notched arrow points referred to by Reeves (1983) as Head-Smashed-In Corner-Notched have been found in several Avonlea assemblages, usually occurring in very low frequency or as single specimens. Avonlea materials are abundant in the project area, with sites consisting of both bison kills and campsites. Well documented Avonlea sites in northern Montana include Timber Ridge (Davis 1966), Three Buttes (Brekke 1969), Lost Terrace (Davis and Fisher 1988), Wahkpa Chu'gn (Davis and Stallcop 1966), Herdegen's Birdtail Butte (Brumley 1990),  Henry Smith (Ruebelmann 1988), Fantasy, Beaver Bend, and TRJ (Tratebas and Johnson 1988). Ceramics are reported only from the Fantasy site.  Avonlea, because of its frequent occurrence and highly distinctive character, has received considerable attention from researchers both within and outside northern Montana (Davis 1988). It is clearly the best documented cultural complex in northern Montana.

The second major projectile point complex recognized here for the Late Prehistoric, is termed the Prairie/Plains side‑notched complex.  Assemblages of this complex are characterized by a dominance of Prairie/Plains side-notched point forms (Brumley and Dau 1988). It should be emphasized that the Prairie/Plains complex encompasses a broad range of cultural variation.

The Protohistoric - Historic Period

One of the first group of explorers to examine portions of northern Montana were Lewis and Clark, who viewed the region along the Missouri River valley in May of 1805 and again in July of 1806 (Coues 1893). Then and throughout the Historic Period, north central Montana was occupied principally by the Atsina and Gros Ventres, and secondarily by the Assinniboine and Blackfoot. Archaeological sites known to date to the Protohistoric - Early Historic period, however, are quite rare in northern Montana. Brumley (1966) and Milne-Brumley (1974) documented historic human burials along the Milk River valley near and north of Havre, and Ann Johnson (1975) described a petroglyph boulder near the Missouri River valley depicting horse hoofprints.

Although archaeological investigations within northern Montana began to increase beginning in the 1960s and have contributed to the development and continued refinement of local and regional culture history models, no record of previous professional cultural resources inventories or other studies exists specifically within the defined Bear’s Paw Battlefield study area were identified (Kurtz 1994). However, Passmann (1990) reports on the results of a negative inventory of a proposed stockwater reservoir immediately outside the defined study.   

Although as yet not fully documented, in 1992 and in several subsequent years former Blaine County museum director Leroy Anderson and amateur archaeologist Paul English personally conducted an informal but thorough surface inventory within much of the Bear’s Paw Battlefield study area. Presently, materials they have identified in that study as being of definite or probable prehistoric age consist solely of surface stone features. Anderson and English have plotted many of the identified surface stone features on field maps. They have recorded the locations of approximately 90 tipi ring sized stone circles; 25 individual stone cairns, and nine stone alignments or drive lines. It should be pointed out that none of the materials identified by Anderson and English have yet been formally recorded as defined sites. Further, complete counts of surface stone features within the project area were apparently not taken. Certain features of a problematic nature were excluded. The information gathered through the efforts of Anderson and English, however, provide the only available dataset specific to the study area from which it is possible to make general statements regarding the kinds and quantities of cultural resources that are or may be present.


Prehistoric and Protohistoric Site Types Found During Inventory

Lithic scatter

Lithic scatters usually consist of limited to extensive quantities of cultural materials largely or totally exposed atop the ground surface.  Observed materials most commonly consists of chipped stone debitage and/or fire cracked rock (FCR) with even more limited quantities of well-made stone tools and tool fragments. Most lithic scatters within the project region are situated on stable ground surfaces or exposed in slightly eroded areas and; in the former case at least - lack stratification, preserved datable organics, and diagnostic artifacts.

A single fire cracked rock with six associated pieces of debitage was found in the park.  The find area was covered in moderate to heavy vegetation and the ground surface could not be fully examined. The few flakes appear to have derived from locally available quartzite and argillite cobbles. Age and function could not be determined.

Stone Circles

Surface stone circles, consisting of locally available unmodified stone, are the most common class of archaeological sites/features found within northern Montana. Typically, stone circles within northern Montana consist of ovate to roughly circular concentrations of large cobble to small boulder size stones.

A total of 36 stone circles were found during the field investigations. These stone circles varied in diameter from 2.5 meters to 6 meters. These were also made up of small to moderately sized cobbles.

During the 2000 inventory 30 wooden stakes were found and recorded that probably relate to the Anderson and English 1992 inventory conducted after a range fire. Anderson and English, based on their maps noted at least 52 stone circles and sixteen cairns in this same area. Dense vegetation prevented the crew from determining if the identified wooden stakes represented stone circles, cairns, or other features.

Cairns and Rock Alignments

Cairns are another type of surface stone feature common within the project region. Rennie and Brumley (1994) define them as small to large piles of cobble to boulder size stones. They state cairns can vary significantly in size consisting of anywhere from 2 - 3 stones, to massive structures containing several tons of stone. Cairns can be found as separate structures, or as elements of larger features such as stone alignments. In addition to having been constructed by historic and prehistoric native peoples, cairns were also constructed by historic Euro-Americans-particularly while clearing cultivated lands of stone. In determining whether specific cairns are associated with the activities of past Native Americans or historic Euro-American activities Rennie and Brumley (1994) define four primary characteristics of a cairn:

1)  the location of the structure in relation to evidence of other prehistoric or historic materials or activities;

2)  the extent of sodding around the stones comprising a cairn;

3)  the extent of lichen cover on the exposed surfaces of the stones comprising a cairn;

4)  the nature of any associated cultural materials

The 2000 field investigations noted and recorded 31 rock piles that meet the criteria of cairns (Figure 6). Two rock piles are most likely rock rubble piles resulting from recent field clearing work.

The remaining 29 are small piles of rocks and cobbles scattered over the. Their age and function could not be reliably determined from the available evidence. Anderson and English plotted at least 32 cairns (Figure 5) in 1991 in the same general areas.

Linear surface stone features consisting of a series of cairns and/or solid lines of rocks are another class of surface stone feature common to northern Montana according to Rennie and Brumley (1994). Such features are generally interpreted as reflecting past communal ungulate hunting activities by Native Americans.

The current project recorded ten rock alignments. Twelve stone alignments have been identified by Anderson and English in the course of their inventory within the project area. Ten roughly correlate with the ten found during the current project work.

The Battle Context -- A Brief History

The final battle of the Nez Perce War of 1877 took place in the valley and on the surrounding terraces of Snake Creek. Detailed accounts of the battle can be found in McWhorter (1986; 1991), Beal (1963), Ege (1982), Brady (1916), Hare (1916), and Greene (1995) among others. The Bear Paw battle, for the purposes of this archeological report, is only briefly summarized.

The Nez Perce after fighting a number of pitched battles and several skirmishes were moving north toward Canada. They halted on Snake Creek, north of the Bear Paw Mountains, in late September to rest themselves and their weary horses. The Nez Perce had outrun General O. O. Howard's command, and believed themselves to be in a position to rest without fear of attack.

The Nez Perce were unaware that Colonel Nelson A. Miles had mounted an expedition to find the Nez Perce, composed of two companies of the Second Cavalry, three of the Seventh, four of the Fifth Infantry mounted on captured Sioux horses, two unmounted Fifth Infantry companies, a  1.65 inch (42mm) Hotchkiss cannon, and a twelve pound Napoleon cannon, as well as an attendant supporting wagon train.

During the morning hours of September 30, Miles began an attack on the Nez Perce village that lay in a shallow crescent shaped valley. High bluffs and terraces dominated the valley. Miles deployed Companies F and H, Second Cavalry to capture the horse herd. These companies moved north and west. Their movements were further west than intended, but they succeed in capturing most of the herd. Company G of the Second pursued some Nez Perce, who escaped the camp, and engaged them about five miles north of the Bear Paw battlefield.

Companies A, D, and K, Seventh Cavalry attacked the village from the south side, but were repulsed with significant losses. Several charges were made by the army during the day. Mounted elements of the Fifth Infantry moved along the village's eastern bluffs and engaged the Nez Perce. Nez Perce losses were about twenty-two killed, although casualties were costly to the small Nez Perce force. The Nez Perce used the terrain effectively and dug riflepits to provide cover for the warriors. The Nez Perce effectively held their ground, giving way slightly to the ever tightening circle of soldiers.

By evening Miles had effectively encircled the village, although White Bird and about 50 other Nez Perce escaped and finally reached haven in Canada. Miles could not destroy or capture the village without additional significant losses. He brought up the artillery and bombarded the camp. Skirmishing from riflepits continued for the next three days. Neither side made any headway.

Howard, with his command, arrived on October 3 and the stalemate was broken. Joseph surrendered his band on October 5. The Nez Perce had twenty-five killed with at least 46 wounded. The army had twenty-three killed and 45 wounded. At least one and possibly two wounded later died. Of the soldier dead all but four were Seventh Cavalrymen. The soldier dead were buried on the field, on the terrace south of the village site. The remains were later exhumed and moved to Custer National Cemetery. The Nez Perce dead were also buried on the field, and probably remain there today. 


Relic Collecting and Metal Detecting at Bear Paw

Relic collecting at the site of the Bear Paw battlefield has been a local recreational activity for at least 40 years according to several local informants, including Andy Anderson, Paul English, Gordon Pouliot, and Seasonal Ranger Jim Magara. Perhaps one of the first persons to undertake examination and documentation of the battle's physical evidence was L. V. McWhorter. McWhorter's interest in the Nez Perce Campaign is legendary. Accompanying battle participants, primarily Nez Perce, McWhorter ventured to the various battlefields of 1877 and recorded the memories of the participants. McWhorter, using wooden stakes, marked many of the sites and features pointed out to him by the then elderly participants (copies of the maps and notes are on file at Big Hole Battlefield and Midwest Archeological Center). In 1935 and 1936 C. R. Noyes mapped the staked locations with chain and transit (Figure 7). Noyes produced the first truly accurate map (a copy of the map is on file at Big Hole Battlefield and Midwest Archeological Center) of the field with the primary landscape denoted as well as the locations staked by McWhorter and the Nez Perce battle participants.

It is well known that the Bear Paw battlefield was a ripe area for finding relics related to the battle. Incidental and serious collectors ranged over the site for many years. Unfortunately, most of these efforts have gone undocumented. Fortunately, there are several notable exceptions.

Thain White visited and extensively researched and collected the site from the late 1950s through at least 1975 (Figure 8). White relied heavily on the Noyes map for his collection efforts. He transcribed Noyes survey fieldnotes (A copy is on file at Big Hole Battlefield and Midwest Archeological Center), thus saving them for current research efforts, as well as transcribing many of McWhorter's notes related to incidents during the battle. Gordon Pouliot (personal communication July 23, 1994) indicated he had metal detected the area with White. Both produced reasonably extensive documentation of their finds (copies on file at Big Hole Battlefield and Midwest Archeological Center). The documentation focuses primarily on artifact descriptions rather than where individual items were found. While this information is not as precise as we might wish it by today's standards, it, nevertheless, remains very good documentation of their collection efforts and finds. This is particularly true of White's collection. White loaned his collection to a now defunct museum in Spokane, Washington. When the collection was transferred to the Cheney-Cowles Museum, and subsequently returned to White, many artifacts were found to be missing.

White also took the trouble to document the O. W. Judge collection of Bear Paw artifacts (copy on file Big Hole Battlefield and Midwest Archeological Center). Judge, another avid collector found the items on the field in 1962. White recorded these finds in the same consistent manner he had recorded his own collecting efforts. Thus White's documentation remains a primary resource in determining the type and quantity of artifacts associated with the battlefield commonly found during the era he collected. The remaining artifacts from Bear Paw are now on loan and displayed at the Blaine County Museum thanks to the diligent efforts of Andy Anderson.

Gordon Pouliot also has an extensive collection of Bear Paw material (Figure 9). It is still in his personal possession. His collection, or at least a portion, is displayed in show cases in an outbuilding on his property. The Pouliot material from Bear Paw duplicated the White collection, although included among the artifacts are a large group of crushed .45-70 cartridge cases, all smashed in the same manner. It is suspected these are army cases probably crushed in obedience of orders to keep the Nez Perce from capturing them for reloading ammunition.

Another collector of the 1960s and early 1970s was Norman Johnson. Mr. Johnson also documented his find locales (copy on file Midwest Archeological Center). Most of his collection now resides in the Blaine County Museum, where it is on display (Figure 10).

Leroy, "Andy", Anderson is another avid Bear Paw researcher and park supporter. He, Paul English, and a few others began a metal detector and visual inventory of the battlefield and surrounding area after a range fire in 1991. They too have documented their find locations (a copy of the notes and map is on file at Bear Paw and Midwest Archeological Center. The map is included as part of the Rennie and Brumley (1994) prehistoric overview. Among their finds are .45-70 cartridge cases and tincans. The cans and cartridge cases where found outside the boundary fence when they were staking the various features found after the fire. Most cartridge cases are 45-70 Benet primed.  Both short and long crimp internally primed cases are present. There are two externally primed with slightly convex heads in Anderson's collection which may post-date the battle. The can tops and one body may be period. They are all hole-in-top with heavy solder. Two are rectangular meat can types, but all are small. The round type may be a size 2 or 2 1/2.


Table 1

 Summary of Artifact Types Collected at Bear Paw Battlefield




Battle-related Features Recorded During the Inventory

The battlefield has a number of currently visible features associated with the battle. Some are marked and interpreted. Riflepits are present and easily seen on all portions of the battlefield.

Fifty-five depressions were recorded during the current project that appear to fit the criteria for riflepits or shelter pits, perhaps more properly called hasty entrenchments. Thirty-two of the depressions or entrenchments were found and recorded in the village area and on the ridges above the village site (Figures 11, 12, 13, 14). Thirteen were recorded in six different locales south and east of the army positions, (Figures 15, 16, 17) and ten were recorded about west of the village area.

These latter ten are the cannon emplacement and associated riflepits used by the artillery crew and its protective force during the bombardment of the Nez Perce village during the battle. The 12 lb Napoleon cannon emplacement is a large pit with eight riflepits located north and south of the large pit. The arrangement is linear.

The thirteen hasty entrenchments found south and east of the village area are uniquely situated. These riflepits are situated so they overlook the logical routes of movement to and from the village. It is likely these depressions are hasty entrenchments placed in strategic locations to thwart any Nez Perce trying to leave or enter the village via cover provided by the coulees.

When mentions of entrenchments occur in the Indian War literature they are often referred to as hastily dug entrenchments, quickly dug, a mound of earth thrown up for protection, or a shallow riflepit. Such statements leave the impression of a haphazard construction to meet an immediate and life-threatening need. These references also convey a feeling of unpreparedness on the part of those constructing the earthwork; a lack of familiarity, training, or knowledge of the purpose or use of a earthwork, beyond that of turning a few bullets in the immediate engagement. None of this could be further from the truth.

The common perception of military engagements in the west is one of a running fight between antagonists or hit and run tactics of Indians versus the Euro-American encroacher. A review of the historic literature relating to the Indian War era demonstrates that various types of earthworks were used in combat situations between Native Americans and Euro-Americans. An interesting sidelight is that Native Americans did construct and utilize several types of entrenchments in much the same manner as the Anglo-American combatants. Limited archeological investigation (Bray 1958; Scott 1994) of earthworks in the trans-Mississippi West demonstrates that the earthworks constructed by Euro-Americans, specifically soldiers, were not hasty or haphazard as is the common perception. They were constructed according to procedure outlined in various military guides of the period.

The American classic, and one that guided the construction of earthworks in the Mexican War and the Civil War, is D. H. Mahan's 1836 A Complete Treatise on Field Fortification, with the General Outlines of the Principles Regulating the Arrangement, the Attack, and the Defense of Permanent Works. Hasty fortifications were defined as those constructed so that troops could take better advantage of the opportunities of natural cover (Mahan 1847). Nevertheless, hasty entrenchments were not to be the rule. American and, for that matter, European military thought was dominated by the concept of massed frontal assault. The use of entrenchments was to play a defensive role.

Dennis Mahan's treatise on field fortifications was uniquely American, in that it recognized most American wars would be fought by militia and only the few regulars would be the most disciplined. If defense was necessary then the militia could build and occupy field fortifications strong enough to resist the enemy's frontal assault until a well organized counter assault could displace them (Hagerman 1965).

Not until the latter part of the nineteenth century did military theorists begin to formalize the concept of small unit tactics. Small unit movement, essentially the squad level, was first introduced in Emory Upton's 1872 Infantry Tactics, but these were not small unit fighting tactics, only mechanical movements.

Unfortunately the U.S. Army published few formal field manuals for small units before the beginning of the 20th century. However, a number of practical guides for officers were privately published throughout the century to bridge the gap left by the lack of official guidance available outside the West Point classroom. One of the most used guides was Mountain Scouting by Captain Edward Farrow. Farrow was an instructor at West Point when he wrote his practical guide in 1881. He had seen active field service during the Nez Perce campaign of 1877 and was with Howard, when he arrived at the Big Hole. Farrow (1881:243) noted "The history of all battles of late years has shown the expediency of making use of natural shelter or constructing field entrenchment's.'' 

Military manuals of the early twentieth century are more structured than Farrow's instructions, but they describe essentially the same procedure for digging riflepits and trenches in the face of the enemy. Moss (1918:385-7) is a good example of such a work, and provides some clear definitions of the purpose of such works.

Moss (1918:385) states the object of field fortifications are twofold; first to increase the fighting power of the troops by enabling the soldier to use his weapons with the greatest possible effect, and second to protect the soldier against the enemy's fire. While the military objective might be stated in that order, the doughboy might have reversed the priority order.

Although written nearly forty years after Farrow's 1881 publication the Moss description of entrenchment methods and types is very similar. It can be argued that riflepits or hasty entrenchments, those meant to be constructed in the face of the enemy, did not change in type or need. Even the World War II "foxhole" as described in the manuals of that period do not differ significantly in purpose or construction from that advocated by Farrow in 1881.

Late Historic Period Artifacts and Features

Two late nineteenth or early twentieth century trash dumps, two dugouts (Figure 3), and a single isolated ceramic fragment were recorded during the field work. The ceramic fragment was found near the park's southeast corner, but outside the boundary. It is a fragment of white ironstone.

A trash dump or trash scatter containing miscellaneous metal and portions of a ceramic figurine were located south of the park boundary fence near the highway, approximately 650 meters south southwest of the monuments.

A second trash scatter associated with two dugouts cut into a ravine sideslope was recorded 530 meters northwest of the monuments. The trash scatter contained a variety of metal and glass. The glass was from bottle and jars. The diagnostic material, scattered over a 50 meter area on the flat above the ravine, dated to the early twentieth century, certainly the first quarter of the century.

One dug out was immediately west of the trash scatter. It was an irregularly shaped, but roughly rectangular depression about 8 meters long east to west and 4 meters wide north to south. This may have been the site of a shed or root cellar. The second depression was cut into a south facing bank of the ravine as it curved from north to northwest. The depression is about 20 meters long east to west and 8 meters wide north to south. This may have been the site of a domestic structure, possibly a house.


General Observations

The Anderson-English work in 1992 recorded 147 possible features exclusive of riflepits (Figure 5). These possible features included 103 stone circles, 32 cairns, and 12 rock alignments. The 2000 archeological inventory recorded 36 stone circles, 29 cairns, ten rock alignments or portions thereof, and 66 wooden stake location (Figure 3). The stake locations had such dense vegetation covering the area that the feature the stake marked could not be discerned. Exclusive of the stake locations and riflepits the 2000 project recorded 75 features. The 2000 project recorded 51% of the features found by Anderson and English after the 1991 range fire. If the stake locations are factored in then the 2000 project found 141 locations, exclusive of riflepits or 96% of the Anderson and English find locations. Since the vegetation was too dense to identify the stake locations in 2000, it more appropriate to say that only 51% of the 1991 features were relocated (Figure 20).

Anderson and English recorded eighteen possible riflepit locations in 1991. They did not record any riflepits in the Nez Perce village area. Instead their riflepit locations are in the southeast area of the park and outside the boundary. The 2000 project recorded 46 riflepits, most of which are in the village area. Excluding the village riflepits, the 2000 project recorded twelve riflepits around the park and east and south of the boundary. While there is some correlation between the two projects' riflepit locations, it is general at best (Figure 20). For instance, the current project did not record as many riflepits near the southeast corner as Anderson and English did, three versus twelve, respectively. The differences may be due to several factors. One is the density of the vegetation change between 1991 and 2000. A second difference is the unwillingness of this author to call every subtle depression a riflepit, if they do not meet any other criteria for hasty entrenchments, such as a definitive earth berm or rocks placed more on the firing side than the other.

C. R. Noyes mapped the L. V. McWhorter stake locations in 1935 and 1936. According to the map, Noyes placed 153 stakes around the field indicating the locations of riflepits, Nez Perce tepee sites, and various other features and locales associated with the Bear Paw battle. The staked locations were replaced by numbered metal capped rebar several years ago. Those metal stakes were cut down and set flush with the ground in 1998. The 2000 visual inventory found only 45 of the Noyes/McWhorter staked sites, or a little over 29%. The thick grass prevented the team from seeing most of the staked locations. If metal detectors had been employed to find the locations, there is little doubt that many more would have been found.


Conclusions and Recommendations

The Bear Paw battlefield is well known and physical evidence of the battle has been collected by the various metal detectorists. Some of these collecting efforts are documented. Copies of various notes and collections descriptions are on file at Big Hole Battlefield, Bear Paw Battlefield, and Midwest Archeological Center. The McWhorter stake locations as depicted on the Noyes map was digitized using AutoCad in 1994 (Scott 1995). The 2000 inventory used a total station transit and a GPS unit to map the locations of all field finds. Using Noyes survey stations, several of which were recorded during the 2000 inventory, as registration points in the AutoCad map the 2000 data was imported to a copy of the 1994 digitized version. The correlation between the points is not perfect. There is a relatively consistent error between the digitized Noyes stake locations and those Noyes stakes mapped in 2000. That error is about 13 meters between the digitized version and the 2000 version. There is a greater error observed between the Anderson-English digitized locations and the 2000 mapped locations of the same features. We believe the error between the digitized Noyes locations and the 2000 mapped Noyes stakes is due to standard error commonly found in digitizing data from photocopied maps rather than the original map, use of the only available USGS 7.5 minute quadrangle maps that are over 20 years old, as well as minors errors on the original map, and minor registration errors in the digitizing. These can be rectified with further detailed mapping of the site at some point in the future.

The differences between the Anderson-English locations and those same features mapped in 2000 are relatively straightforward. Anderson and English used the available orthophoto map of the area on which they plotted their finds sites by eyeball. Without clear and easily defined reference points some error in plotting occurred. The differences in the 1991 data and the 2000 mapped data are noticeable but not extreme. The 2000 mapped data are the more precise.

Regardless of accuracy in the 2000 mapping, one is struck by the significantly lower numbers of features found and recorded by the current effort than those of previous efforts. Compared to the Anderson-English 1991 work, the 2000 team only found 51% of those features. Even allowing for inter-observer error, this is a significant difference. While the 2000 team did find 96% of the Anderson-English staked locales the dense grass growth since 1991 obscured our ability to determine what was at most of those staked locations.

The effect of the dense grass growth is doubly noted when comparing the known Noyes stake locations versus those recorded in 2000. Only 29% could be found by visual inventory. The dense vegetation matte confounded our ability to effectively find the previously known sites and features. We estimate that the 2000 inventory found and recorded, at best, only 30% to 40% of the known sites and features within the Bear Paw GMP boundary. For this reason the 2000 inventory effort can only be considered a reconnaissance level at best. It is also for this reason that no site forms were generated for this project. Since only 30% or 40% of the known features were relocated it was determined that any site forms generated would be full of error and imprecision that would bias future recording efforts. The data generated for this project should be considered an initial recording effort and be used as baseline data for future inventory and recording efforts.

One recommendation resulting from this inventory is that detailed mapping of sites and features of the Bear Paw battlefield should continue as opportunities permit. The 2000 inventory and mapping is incomplete due to the thick vegetation matte. As prescribed burns are done or the vegetation matte is reduced by other means, mapping and feature recording should be done as soon as possible after the event.

At least one conclusion can be drawn from the previous and current work concerning the distribution of hasty entrenchments or riflepits. It is patently obvious that some riflepits were placed at strategic positions overlooking several deep coulees that could be used for escape or other movements. The shape of the riflepits found in these locations is relatively regular and the orientation is generally toward the Nez Perce positions. It is reasonable to assume that these are army riflepits placed by design to prevent the Nez Perce from escaping the cordon that was thrown up around them during the battle. As such their placement is the physical reality of and provides some additional insight into the strategy and tactics employed by Nelson Miles and his officers.

Nez Perce and Army riflepits appear different in construction and design. In interpreting cultural differences in warfare the riflepits offer a unique opportunity to explain what constituted appropriate cover for combatants reared in different cultures. Several riflepit features are readily visible and conveniently located adjacent to the current interpretive trail. These features could be excavated to ascertain their construction techniques. They could be compared to one another to determine cultural differences in construction and further compared to riflepits from other battles to expand our understanding of how different cultures built and used riflepits.

The Bear Paw battlefield has yielded many artifacts to collectors and earlier researchers. Those extant documented collections form a core of physical evidence on the battle. They also demonstrate the research potential of the in situ artifacts. The battlefield has yielded and undoubtedly still contains patterned data that can be extracted, analyzed, and interpreted using available archeological techniques and theory. The battlefield should be systematically metal detected to recover that patterned information at some point in the future when all parties can agree to the procedure. The in situ data coupled with additional analysis of the extant collection has a very high potential to reveal details of the battle not recorded in the historic record, and to yield new insights about the battle and its participants


References Cited

Beal, Merrill 1963 "I Will Fight No More Forever". Ballentine Books, New York.

Brady, Cyrus Townsend 1916 The Epic of the Nez Perce. pp. 3-43, in Northwest Fights and Fighters, edited by Cyrus Townsend Brady, Doubleday, New York.

Bray, Robert T. 1958 A Report of Archeological Investigations at the Reno-Benteen Site, Custer Battlefield National Monument June 2 - July 1, 1958. Ms on file Midwest Archeological Center, Lincoln.

Brekke, A. 1969 Antiquity in North Central Montana. Archaeology in Montana 10(3):1-11. 1970


Three Buttes Bison Kill (24BL104). Archaeology in Montana 11(4):39-54. Brumley, J.H. 1966


The Lost River Burial. Archaeology in Montana  7(4):1-14. 1975 Radiocarbon Dates from the Wahkpa Chu’gn Site. Archaeology in Montana 16(2):105-116.

1978 McKean Complex Subsistence and Hunting Strategies in the Southern Alberta Plains. Plains Anthropologist Memoir 23(82, Part 2):175-193.

1988a An examination of archaeological materials from the Bear Paw Mountains area of northern Montana, 2 Vols. Consultants report (Ethos Consultants Inc., Havre, Montana)

prepared for and on file with Montana State Historic Preservation Office Helena, Montana.


1988b Medicine Wheels on the Northern Plains: A Summary and Appraisal. Archaeological Survey of Alberta Manuscript Series. Alberta Culture, Edmonton. No. 12.


1990 Investigations at Herdegen’s Birdtail Butte site (24BL1152). Consultants report (Ethos Consultants Inc.)


nd Description and evaluation of various ceramic materials from various areas of Montana. Manuscript in possession of the author.


Brumley, J.H. and Barry J. Dau 1988 Historical Resource Investigations Within the Forty Mile Coulee Reservoir. Archaeological Survey of Alberta Manuscript Series No. 13. Alberta Culture, Edmonton.


Brumley J.,  and P. Rennie 1993 The results of investigations at the King site along the east margins of the Little Rocky Mountains. Consultants report prepared for the Fort Belknap College, by Ethos Consultants Inc. Havre, Montana.


Coues, E. 1893 History of the Expedition Under the Command of Lewis and Clark. Dover Publications Inc. New York (Reprinted 1965)


Davis, L.B. 1966 Avonlea Point Occurrence in Northern Montana and Canada. Plains Anthropologist 11(32)100-116.


1988 Avonlea Yesterday and Today: Archaeology and Prehistory. Saskatchewan Archaeological Society.


Davis, L.B. and J.W. Fisher Jr. 1988 Avonlea Predation on Wintering Plains Pronghorns. In

Avonlea Yesterday and Today: Archaeology and Prehistory. L.B. Davis editor. pp. 101-118.


Davis, L.B. and E. Stallcop 1965 The Keaster Site (24PH401): A Stratified Bison Kill Occupation in the Missouri Breaks Area of North Central Montana. Montana Archaeological Society. Memoir No. 2. 


Deaver, Ken 1980 Archaeological site distribution in north Blaine County, Montana. Consultants report (Pro-Lysts Inc.) for the Bureau of Land Management, Lewistown District Office, Eugene, Oregon.


Dyck, I.G. 1983 The Prehistory of Southern Saskatchewan.  In Tracking Ancient Hunters: Prehistoric Archaeology in Saskatchewan, edited by H.T. Epp and I. Dyck, Saskatchewan Archaeological Society, Regina.


Ege, Robert J. 1982 After the Little Bighorn, Battle of Snake Creek Montana Territory, September 30- to October 5, 1877. Fred Werner, Greeley, Colorado.


Farrow, Edward 1881 Mountain Scouting. Privately published by the author, New York.


Forbis, R.G. Nd Signal Butte and MacHaffie: two stratified sites on the northern Great Plains. Unpublished manuscript on file, Department of Archaeology, the University of Calgary.


Frison, George C. 1991 Prehistoric Hunters of the High Plains. Academic Press, New York.


Greene, Jerome A. 1995 The U.S. Army and the Nee-Me-Poo Crisis of 1877: Historic Sites Associated with the Nez Perce Conflict. Draft Report, Pacific Northwest Regional Office, National Park Service, Seattle, Washington.


Hagerman, Edward 1965 The Evolution of Trench Warfare in the American Civil War. Unpublished Ph.D. dissertation Department of History, Duke University.


Hardin, Albert N. and Robert W. Hedden 1973 Light But Efficient A Study of the M1880 Hunting and M1890 Intrenching Knives and Scabbards. Taylor Publishing, Dallas.


Hare, W. H. 1916 Chief Joseph's Own Story. pp. 44-75 in Northwest Fights and Fighters, edited by Cyrus Townsend Brady, Doubleday, New York.


Haury, Cherie E. 1989 The North Dakota Archeological Association State Centennial Project: Fort Dilts (32BO6). Department of Anthropology Contribution No. 252, University of North Dakota, Grand Forks.


Johnson, Ann 1975 Hoofprint boulder (24RV1026). Archaeology in Montana 16(1):43-48.

Keyser, J.D. 1986 The Evidence for McKean Complex Plant Utilization. Plains Anthropologist 31(113):225-235).

Kurtz, L. 1994 Results of site file search request to Patrick Rennie, Ethos Consultants Inc., from Leanne Kurtz, Montana State Historic Preservation Office, Dated April 29, 1994.


Lobdell, J. 1973 The Scoggin Site: An Early Middle Period Bison Kill. The Wyoming Archaeologist. Vol.XVI, No.1.


Mahan, Dennis 1847 An Elementary treatise on Advanced-Guard, Out-Post, and Detachment Service of Troops. Wiley and Putnam, New York.


McWhorter, L. V. 1986 Hear Me My Chiefs! Nez Perce History and Legend. Edited by Ruth Bordin. Caxton Printers, Caldwell, Idaho.


1991 Yellow Wolf: His Own Story. Caxton Printers, Caldwell, Idaho.


Milne Brumley, Laurie. 1974 The Fox Burial Site (24HL413). Archaeology in Montana  15(2):1-26.


Morris, E., R. Blakeslee, and K. Thompson 1984  Preliminary Description of McKean Sites in Northeastern Colorado. In McKean/Middle Plains Archaic: Current Research. M. Kornfeld and L. Todd, eds. pp. 45-50. Occasional Papers on Wyoming Archaeology No.4.


Moss, James 1918 Manual of Military Training. George Banta Publishing, Menasha, Wisconsin.


Munson, G. 1990 Archaeological investigations at 24RB1164. Report prepared for Peabody Coal Company’s Big Sky Mine, Area B Rosebud County, Montana, by GCM Services, Inc., Butte, Montana.


Passmann, D. 1990 A surface inventory of approximately 1 acre of land. The inventory was conducted in response to a proposed stockwater reservoir immediately outside defined study area.


Reeves, B.O.K. 1969 The Southern Alberta Paleo-Cultural-Paleo Environmental Sequence. In Post-Pleistocene Man and His Environment on the Northern Plains, edited by L.B. Davis, R.G.

Forbis, O.A. Christensen and G. Fedirchuk, pp. 6-46. The University of Calgary Archaeological Association and the Students’ Press, Calgary.


1973 The Concept of an Altithermal Cultural Hiatus in Northern Plains Prehistory. American Anthropologist 75(5):1221-1253.


1983 Culture Change in the Northern Plains: 1000 B.C. - A.D. 1000. pp. 390. Archaeological Survey of Alberta Occasional Paper No. 20. Alberta Culture, Edmonton.


Reher, C. 1979 The western Powder River basin survey. Unpublished report prepared for the Bureau of Land Management, Wyoming State Office, by the University of Wyoming.


Rennie, Patrick and John Brumley 1994 Prehistoric Archeological Overview and Assessment of Bear Paw Battlefield State Park, Blaine County, Montana. Ethos Consultants, Havre, MT, prepared under NPS PO for Midwest Archeological Center, on file, Midwest Archeological Center, Lincoln, NE


Rice, E. A. 1874  A System of Rapid Field Intrenchments for Infantry. Washington, D.C.


Rossillon, M. 1985 Preliminary report of archaeological and historic investigation in the Fort Belknap-East Highway project area (F1-7(4)430). Report on file, Montana Department of Highways, Helena.


Ruebelmann, G.A. 1983 An Overview of the Archaeology and Prehistory of the Lewistown BLM District, Montana. Archaeology in Montana 24(3)1-165.


1988 The Henry Smith Site: An Avonlea Bison Procurement and Ceremonial Complex in Northern Montana. In Avonlea Yesterday and Today: Archaeology and Prehistory, edited by L.B. Davis. pp. 191-202.


Scott, Douglas D. 1994 A Sharp Little Affair: The Archeology of the Big Hole Battlefield. Reprints in Anthropology Volume 45, J and L Reprints, Lincoln.

1995 Historic Archaeological Overview and Assessment of the Bear Paw Battlefield, Blaine County, Montana. On file Midwest Archeological Center, National Park Service, Lincoln, Nebraska.


1997 An Assessment of the White, Pouliot, Johnson, and Anderson-English Relic Collection from the Bear Paw Battlefield, Blaine County, Montana. On file Midwest Archeological Center, National Park Service, Lincoln, Nebraska.


Tratebas, A. and A.M. Johnson 1988 Three Avonlea Bison Kill Sites in the Milk River Drainage of Northeastern Montana. In Avonlea Yesterday and Today:  Archaeology and Prehistory, edited by L.B. Davis. pp. 89-100.


Tutherly, Herbert E. 1898 Elementary Treatise on Miltary Science and the Art of War. Free Press Association, Burlington, Vermont.


Vickers, R. 1986 Alberta Plains Prehistory: A Review.  Archaeological Survey of Alberta Occasional Paper No. 27. Alberta Culture, Edmonton.


Wheeler, R.P. 1952 A Note on the “McKean Lanceolate Point”.  Plains Archaeological Conference Newsletter 4(4):45-50.

bottom of page