Definition of Warfare and Its Symbolism

Within this volume warfare is broadly considered as a complex mix of ritual warfare, territorial disputes, plunder, and captive-taking for trade and for rebuilding local populations following epidemics (Clark and Sundstrom 2010). The dictionary considers a wide range of definitions for “warfare,” but the common denominator is intergroup conflict or struggle of any kind. Nowhere do dictionary definitions specify the number of people involved, kinds of weapons, kinds of captives, or length or intensity of the conflict.

Anthropological studies of warfare focus on why people go to war, benefits to the group, how the group is organized, and what weapons and military tactics are employed (Otterbein 2009:4). Warfare is viewed as group action rather than as individual action, with the target being group members rather than particular individuals. Otterbein identifies the goals for uncentralized political systems engaging in war as “defense-revenge, plunder, and prestige” (Otterbein 2009:4). Thus, the wide variety of physical conflicts recorded in historical documents for the northwestern Plains, most of which involve small groups attacking other small groups, all fall within the generalized “warfare” classification. The goals of such skirmishes during the Contact period include all of those identified by Otterbein, although not all for any one battle.

Warfare images in rock art are assumed to be representational and easily recognizable, so we complacently believe we know which images portray warfare and can consider individual figures and scenes within variable cultural contexts. But this is not always the case. It has been pointed out by Chippendale (2009) that before deciding if rock art portrays warfare we must separate warfare from other kinds of physical or spiritual conflict, especially on a personal or interpersonal level. Examples are ritual reenactment (especially in dance), copying conflict postures in social dance or exercise (such as karate or capoeira), competitive games, and even hunting. All functions can be portrayed in similar ways, so the researcher must look for indirect evidence of warfare since physical posture alone may be misleading. Chippendale advocates identifying defensive weapons (e.g., shields, fending sticks) rather than offensive weapons (such as bows, arrows, spears, lances, clubs, hatchets, swords, and guns) as important in deciding whether warfare is being portrayed. Candace Greene’s (1985) recognition that there are rules for reading a warfare scene provides another contextual evidence check. Although not every drawing follows the rule that a conflict scene is read right to left (subject-action-object or in warfare terms—hero, what he did with what weapon, enemy), starting with this concept can help determine whether or not the function of the panel is to relate warfare activity.

When discussing warfare on the northwestern Plains, the distinction between actual fighting and rituals associated with fighting is blurred because of the cultural context of these activities; and since the distinction was not made in the lives of those people, warfare on the Plains usually considers actual fighting and ritual portrayal as the same, as they are viewed here. Rituals, whether portrayed as occurring before or after a fight or as associated with fighting, such as the Sun Dance, are not the same as fighting, although they can sometimes substitute. Likewise, portraying activities such as counting coup (striking an enemy either living or dead with a stick, quirt, bow, or similar object during battle), preparing for a battle by drawing a shield image on a rock wall for power, or drawing one on the wall after the battle to record one’s success, may not be a distinction that is needed by the people in societies where warfare and ritual are intertwined or by researchers attempting to understand how warfare changed through time. However, we do not know for certain that actual and ritual warfare were closely related through time, so separating the two concepts should be attempted whenever possible for the best understanding of how warfare was portrayed in rock art and by whom. On the northwestern Plains there are a few examples of rock art that previous researchers have identified as showing fighting postures and portraying social dance, and in some cases these social dances are directly associated with warfare. The most obvious example of a scene with this function is at the Joliet site (24CB402), with a portrayal of the Grass or Hot Dance conducted by the Hidatsa and their northwestern Plains relatives, the Crow (Keyser and Cowdrey 2008; McCleary 2008a:44–45). At this panel three dancing warriors are carrying a gun, a bow and arrow, and a feather-decorated coup stick as part of a ritual battle (figure 2.2). On this same panel is a woman interpreted recently by Crow informants as having been stolen from another tribe and then thrown away as part of this dance ceremony, representing another aspect of warfare (McCleary 2008a:45).

Figure 2.2. Warfare dance and capture scene at the Joliet Site (24CB402) in Montana. Drawing by James D. Keyser, from Keyser and Cowdrey 2008:26 [figure 7].

On the southwestern periphery of the northwestern Plains, at the La Barge Bluffs site (48LN1640) in southwestern Wyoming, two scenes have been interpreted as rituals associated with warfare. Keyser and Poetschat (2005:67–68) hypothesize, based on ethnographic accounts of Northern Shoshone by Lowie, that one scene portrays coup on a captured woman in front of a line of people as she is adopted into the capturing group, thus portraying a war-related activity but not actually showing warfare (figure 2.3a). A second scene at that site shows a warrior brandishing a pistol and riding in front of a group of people in what the authors consider a celebration of warfare, but again not actually portraying war (figure 2.3b). The audience in both rituals is interpreted either as participants (dancers, celebrants) or simply as observers. These cases support the fact that the context of a single image, including details of its depiction, or the context of a complete scene is critical in determining whether warfare is the theme.

Figure 2.3. Ritual warfare at the La Barge Bluffs site (48LN1640) in Wyoming. (a) Counting coup on a captured woman (bottom right) in front of a group. (b) Mounted warrior riding with a pistol in front of a group.

Common symbols that depict warfare are weapons (figure 2.4a–b), shields (figure 2.4c), armor, fighting posture(figure 2.4c), and people in dominant positions facing opponents in subservient positions. Nothing is more conclusive than scenes showing attacks (figure 2.5a–b) or other warrior activities, such as horse stealing. However, although often considered characteristic of regional rock art, the occurrence of action-showing battle scenes is limited relative to static portraits of humans that represent warriors. When shields are portrayed with weapons, they are usually considered conclusively warfare related, but when a person (usually male) is shown with a weapon and no shield, unless he is portrayed in a battle scene, there is no reason to prefer warfare over hunting, indication of status, or some other message. Shields in non-combat scenes, or even static poses without weapons, are usually assumed to be warfare related, such as those at Bear Gulch discussed by Keyser (chapter 3, this volume). But there are examples in northwestern Plains rock art in which the shield appears to have no association with warfare and may instead portray medicine shields, as symbolic weaponry. We have previously suggested, based on panel context, that the shield may be a personal identifier or have a spiritual connotation, such as assisting in safe passage into the next world (Greer and Greer 2003). Other indicators of warfare include warriors holding severed heads, mounted warriors taking pedestrian captives, and armored people and horses. We also believe that people pierced with spears and arrows, usually interpreted as representing personal injuries through fighting, especially in earlier pre-Contact cases, may not represent fighting but instead may be stylized representations of a different but as-yet unidentified function. Thus, when spears and arrows (from Archaic to Historic) are shown recurringly penetrating specific parts of the body, such as the neck (at an angle), waist (figure 2.4b), knees, and ankles, or even lining the torso, they may not be referring to actual wounding by another individual.

Figure 2.4. Common rock art depictions of warfare: (a) shield-bearing warrior (Carboni site, 24CB404); (b) person stuck with arrow (Recognition Rock, 24RB165); and (c) battle scene (No Water Petroglyphs, 48WA2066).

Figure 2.5. General rock art battle scenes with horse-mounted warriors fighting pedestrians: (a) at White Mountain Petroglyphs (48SW302), and (b) at the Gumby site (24GV139). Some images highlighted with Adobe Illustrator.

Handprints are found throughout world rock art and are especially common in central Montana (Greer and Greer 1999). There are many explanations for their existence, but the prints (whether positive stamps or negative stencils) generally are not considered directly associated with warfare on the northern Plains. Historical documents, however, suggest that single hands or hands associated with warfare scenes may have a different meaning:

In all cases, it is again context that indicates whether warfare is the theme associated with the images. Associated dress, accoutrements, and posture are important to the warfare function interpretation, especially on non-scene, static figures. We have the added benefit that Contact-period warfare on the northwestern Plains has a rich historical record written by people living in the midst of that cultural change.

Weapons in Rock Art

Weapons are important in evaluating depictions of warfare, and we have previously quantified weapons recorded in Montana and Wyoming rock art for types of weaponry relative to function (Greer and Greer 2008a, 2008b). Our analysis (updated in 2013) focused first on weapons in Montana, distribution of weapon types, and likely function relative to the overall scene. Unexpectedly, the rock art mostly suggested changing use of environmental settings through time by different populations. We then expanded analysis to Wyoming, focusing on the northern and eastern parts of the state, that is, the northwestern Plains and Rocky Mountain geographical areas, and excluding the southwestern area with stronger Great Basin cultural associations and Great Basin environment. Based on a sample of nearly 1,000 sites (654 in Montana and 337 in Wyoming) it was found that recorders more frequently recognized—or focused their attention on—weapons than on distinctions in other images, at least to the level of general class, such as bow and arrow, lance, or gun (table 2.1). During our review of these sites we were able to identify 511 weapons (382 in Montana and 129 in Wyoming, or 75% and 25%, respectively).

Shields are usually easily identified, and shield-bearing warriors are often portrayed with active weapons such as oblong rounded-end clubs, pointed elk-tine clubs, and lances (figure 2.6). Bows are shown alone and with arrows, and arrows are depicted with triangular arrowheads and feather fletching, or with fletching only and no point. Arrows are sometimes in the hands of humans and not accompanied by a bow, some are in quivers on people’s backs, some are in flight, and others are shown sticking into shields, animals, or humans. Other weapons include hatchets, guns, lances or spears, and the extremely rare atlatl.

Figure 2.6. Shield warriors with weapons at the Bear Gulch site (24FR2): (a) oblong rounded-end club, (b) elk-tine club, and (c) lance.

In Montana and Wyoming bows are most frequently shown in hunting and warfare scenes, thus indicating what we recognize as a progressive increase in weapon images and probably a gradual change in importance from hunting portrayal to interpersonal conflict from the Late Prehistoric (ca. ad 500) to the Historic (ca. ad 1700) periods. The bow and arrow usually are not in ceremonial rock art scenes, which may be both a temporal and functional distinction. However, Francis and Loendorf (2002:117) discuss ceremonial uses of the bow as portrayed in some Dinwoody sites in western Wyoming to indicate power and association with evil activities, such as shooting people with invisible arrows to cause illness. Thus, like the shield, weapons may not be depicting warfare and instead may have a completely different referent.

Armored horses occur in rock art across both states, with the greatest concentration along the Musselshell River in central Montana (figure 2.7). Of three recorded armored horses on the Wyoming plains, two are pierced with arrows or lances. Of nine recorded in Montana, only one (at the Nordstrom Bowen site, 24YL419) is pierced, and it is the most attacked armored horse on the Northern Plains from Alberta to Colorado. Five of the 12 armored horsemen have associated lances or spears. Even though horse armor may have been designed principally for battle protection, only just over half the images are associated with weapons, suggesting armor may have had other functions on the Plains, such as environmental protection from brush (or thorny plants in the south) but more likely from cold temperatures by allowing retention of body heat. Only one known figure—a pedestrian—is in full body armor (figure 2.8), and he is engaged in conflict with a person on an armored horse, on a site along the Musselshell River.

Figure 2.7. Five of seven armored horses reported from the Musselshell River area. Sketches by John and Mavis Greer and James D. Keyser.

Figure 2.8. Pedestrian warrior and armored horse with rider (Goffena site, 24ML408).

Musselshell Rock Art

The Musselshell River in central Montana was one of the last places tribes could continue their cultural practices of hunting and warfare. The valley lies between the Missouri River to the north and the Yellowstone River to the south (figure 2.1). In 1875, Yellowstone Kelly described the area as “a veritable hunters’ paradise for game of all kinds, including elk, deer, and mountain sheep, and cinnamon, black, and brown bear . . . [and] a good country to run into war parties of the Sioux, Crow, and Blackfeet tribes” (Quaife 1973:117). However, these three cultural groups were not alone. At the mouth of the Musselshell, C. W. Lee, a young man trained in gun repair who lived at the confluence of the Musselshell and Missouri rivers, observed the following tribes between 1868 and 1872 (Hampton 2011): Arapaho, Assiniboine, Blackfeet, Crow, Flathead, Gros Ventre, Pend d’Oreille, Piegan, and Sioux (Santee, Teton, and Yankton). Of these tribes, those most mentioned were Arapaho, Crow, Gros Ventre, and Sioux. In 1878 and 1879 Andrew Garcia reports encountering Assiniboine, Blackfeet (mainly Piegan, but also Blood and Blackfoot), Cree, Crow, Gros Ventre, Nez Perce, Pend d’Oreille, Sioux, and Spokane (Garcia 1967). He wrote that the Musselshell country drew western as well as northern tribes because bad weather in Alberta drove the buffalo south to winter there and provided an ample supply of food for the many visiting groups. This popular wintering location provided opportunities for intertribal conflict but also complicates assigning cultural affiliation to rock art in the valley.

Sites along the Musselshell mostly contain Contact-period rock art, but only about a third appear to be associated with warfare. There has been little archaeological survey here, but 16 rock art sites have been recorded overlooking the Musselshell River, and 31 for the entire drainage, undoubtedly a small percentage of sites actually along the sandstone-rimmed valley. Of the 16, eight have scenes that portray warfare.

The Gumby Site (24GV139) is one of the smallest with a battle scene. A single rider on a horse appears to be leading two riderless horses and shooting a gun toward a pedestrian shown only from the torso up and carrying a bow (figure 2.5b). Other images at the site are of red paint, and based on our prior studies of central Montana chronology (Greer 1995) probably date much earlier than the Contact period but are too deteriorated to be identified.

The Five Guys Petroglyph (24ML394) has two horseback riders carrying long lances, possibly coup sticks, following five humans. One rider has flowing long hair, which differs from the round heads of the rectangular-bodied pedestrians, who have no arms. No other figures have been noted here.

The Rockshelter Shield site (24ML507), the Horned Headgear site (24ML508), and the Musselshell site (24ML1049) have recently been recorded in detail and found to have warfare imagery (Loendorf 2012). The Rockshelter Shield site has several static-pose, shield-bearing warriors, although only one has a clear weapon. The Horned Headgear site has an action battle scene, which is typical of those found at Writing-on-Stone in Alberta. The Musselshell site has several shield-bearing warriors but no active battle scenes.

The other three sites with battle scenes all have armored horses (figure 2.7), although not all are within conflict compositions. The Goffena site (24ML408) has a painted armored horse confronting an armored pedestrian (figure 2.8). The scene is not only unique among armored-horse depictions along the Musselshell, it is also unlike any others on the northern and central Plains and not just because it is a painting rather than a petroglyph. The Goffena horse has a scalplock hanging from the bridle bit, and a rayed headdress, which is referred to as a horse bonnet (Keyser 2012), and neither of these occurs on other known armored horses. The shielded warrior riding the horse has a horned headdress, carries a flagged lance, and has what appears to be a thrusting spear pointed at the pedestrian warrior. The body-armored pedestrian has a horned headdress, carries what may be a flagged coup stick with an attached scalp, and is protected by a large shield in addition to the armor.

The West Ryegate (24GV191) armored horses are all easily recognizable, but no two are exactly alike (figure 2.7). They are spaced along a quarter-mile of bluff, with no two images together, suggesting they were drawn by different people. Horse 1 (AH1) is flanked by a shield bearer to the right and another to the left. The horse armor is a typical triangular skirt with a curved bottom, and a collar to cover the horse’s neck. The horse’s head has been lost to calcium carbonate deposits, but a group of lines just out from the collar suggests a decorative bridle. The shield-covered rider at the top opening of the armor is not detailed. However, there is a deliberate slash across the face (the recurring scar-face motif), one arm and hand, and a suggestion of reins. In front of the rider is a long, vertical lance with a tassel extending from the top, perhaps representing a scalp. Far to the right (not shown in figure 2.7) is a pedestrian warrior, and closer to the left is another warrior facing left away from the horse. This scene does not clearly depict the armored horse interacting with the pedestrians as it does at other sites.

The second armored horse (AH2) at West Ryegate is very large relative to others of this kind. Although dense carbonate deposits surround the incised figure and cover part of the rider, it is still possible to see that the horse armor has broad slightly expanding stripes that form a pattern similar to those at Goffena. There is a tassel off the end of the nose, perhaps a scalp or ring bit chains (or coscojos). Reins end at a lance extending frontward from the shield-bearing rider. Attached near the front of the lance are what may be feathers or a scalp. The warrior has a wide neck and what appears to be a single feather or ponytail extending from the oval head. Protruding from his shield, above the lance, is a plain arrow or another lance.

Horse 3 (AH3) at West Ryegate has typical triangular-shaped armor with an opening in the top for the rider and a collar to protect the horse’s neck. Triangular designs on the armor body may be highly stylized feathers or a pattern in the leather indicating construction. A column of five large dots decorates the front of the armor. This decoration is not on any other recorded armored horse. The pointed-head rider is mostly outside the horse armor but is protected by a personal shield. Lines extending out from the shield on the edge opposite the reins may be from a weapon now not discernible. Superpositioning of the scratches shows that the large shield to the right of the rider and at the top of the armor was engraved before the horse; so presumably the order of engraving was the shield first, followed by the horse armor and horse, and finally the human rider.

Armored horse 4 (AH4) at West Ryegate is on a busy panel also containing at least one horse without armor and rider. Based on superpositioning, the armored horse was not the first of the figures to be incised. The armor is without decoration, apparently to allow the underlying unarmored horse and rider to show through—that is, the engraving order is the unarmored horse and rider, and then the armored horse and rider. The armored rider lacks detail, but the generally rounded body suggests a shield, while a distinctive lance with dangling feathers or scalp protrudes from the back of the armor. The rider of an underlying armorless horse also carries a lance with a possible scalp or feathers.

Although none of the West Ryegate horses is in a definite battle scene, all are associated with weapons or war trophies (i.e., scalps) suggesting they all represent warfare-associated activities. This long bluff also has two other small scenes possibly associated with conflict. One may depict a horse-stealing event—a horse with a down-turned head is partially superimposed onto a conical tipi, and the two are covered with horse prints. The other panel has a well-executed horse with a shielded rider holding a lance or coup stick and being bombarded with arrows from an unseen source.

Two armored horses have been identified at the Twenty-one Guns site (24ML398). Like West Ryegate, this is a large site with several weapons depicted, mostly guns. Also here are several unarmored horses and some shield-bearing humans (one with guns). Neither horse has associated figures, weapons, or rider. There are no indications that the two horses were made by the same person or by any of the artists of the other five armored horses along the Musselshell.

The Musselshell sites contain several hundred elements of Contact-period rock art, and about a third may be associated with warfare. Ethnographies show that tribes in this area had a social structure with status dependent on military achievement (Lowie 1963:114–123), but historical documents often provide a different view of conflict. In the 1800s horse stealing was still the main way to increase status within most northwestern Plains tribes, and historical documents cite this as the main reason for warfare in the Musselshell area (e.g., Garcia 1967; Hampton 2011; Quaife 1973). However, squeezing so many tribes into the small valley because of diminishing buffalo herds, increasing Euroamerican settlements in surrounding areas, and constant pressure from the US military was causing increased skirmishes between small parties (McGinnis 1990). These conflicts arose from too many people using a more constricted space for activities that previously encompassed massive areas.

Ethnohistoric Warfare: What Will We See in Rock Art?

By the mid- to late 1800s, traders and trappers, such as Andrew Garcia, a trader from the border area of Texas in the late 1870s, were living with and marrying into tribes that lived in the valley, and some, like Garcia, were writing extensively and in detail about their time on the Musselshell. The Musselshell valley was not only a place for many tribes to winter, hunt bison, and interact, both in conflict situations and at social events that centered on gambling (Garcia 1967:170, 185), but also for white traders and trappers to intermingle with the Indians or to enter the area as part of the US military, and in some cases both (Quaife 1973). From Garcia we see the same tribes that often skirmished would get together for social parties that lasted for days. While living in a camp of Pend d’Oreille and planning to marry a Nez Perce woman living among them, he witnessed such a party in 1878. Tribes came to gamble at the Pend d’Oreille camp: “Assiniboines and Crees, Bloods, Gros Ventres and Piegans” (Garcia 1967:185). However, peaceful interactions are not common topics in historical documents for the Musselshell area, whereas warfare between the tribes, and later between the tribes and the US military, are much more popular subjects. We also learn from individuals who integrated themselves into the tribes, that Indian groups, even when there was no formal social event, were generally composites from different tribes. Although they recognized tribal distinction among themselves, affiliation would not be obvious to a casual outside observer. For example, Garcia wrote about one camp, “some Spokanes were with them, but most of the band were Pend d’Oreilles from the Kalispell Valley [over 200 mi to the west]. They were camped about three miles from where I was . . . They had come over the year before and had hunted buffalo in the Musselshell country the previous winter” (Garcia 1967:113). He also noted that Indians from west of the mountains generally stayed two or three years before returning home.

Among the conflicts described, those associated with horse stealing are most common. By the late 1870s, stealing horses not only brought prestige within the tribe but increased tribal assets for trade. Horse stealing between tribes and from Euroamericans in the area was a constant in the region (Hampton 2011; Robison 2013). Garcia speaks particularly of Crow and Piegan war parties stealing horses back and forth (Garcia 1967:31, 49, 66). Because of the abundance of horses, Garcia was usually not interested in trading for horses, but he noted that other whites in the area were. Those in the small settlement of Fort Musselshell at the confluence of the Musselshell River with the Missouri fueled horse stealing by offering whiskey in trade: [December 27, 1869] “The Grovents are still here and doing considerable trading: horses and robes for whiskey, although there is a heavy penalty against it, there are plenty that will trade it to them” (Hampton 2011:63). Horse stealing is portrayed in rock art throughout the northern Plains, so to find only one horse stealing scene in the rock art of the Musselshell is surprising, considering its numerous references in the historical record. If cases of horse stealing have the same proportions in rock art elsewhere in the northern Plains, it seems that this was not a war exploit that was commemorated with a frequency relative to how much it was occurring.

Capturing or stealing women during conflict was widespread across the region, but historical references to this for the Musselshell area are few. Keyser, Sundstrom, and Poetschat (2006) reported on the occurrence of women in war and noted only 24 rock art scenes at 16 sites on the entire northwestern Plains that depicted women being captured (figure 2.9). Subsequently, at the Bear Gulch site (24FR2) in central Montana, at least five women were found to be in coup-count scenes and tallies (Greer and Keyser 2008:97–98; Keyser, chapter 3, this volume). In 2009, a woman-capture scene was recorded at the No Water site (48WA2066) in northwestern Wyoming (Keyser and Poetschat 2009:13, 83–91). However, even with the addition of the new panels, capturing or stealing women does not appear to be a popular topic for rock art.

Figure 2.9. Woman-capture scene (with armored horse) at 38HN210, South Dakota.

No woman-capture scenes have been found in the Musselshell drainage, but there also are almost no indicators of gender. Even historical references to stealing woman for the Musselshell are few. There is a general comment that the Blackfeet steal robes in raids, like they do horses and women (Garcia 1967:163), and a specific instance regarding a Blackfeet raid states that

Full-time residents of Fort Musselshell at the mouth of the river were few in number (in April 1870 there were 13 men, 4 women, and 2 children), but there were always many visitors, among them captive women.

Seldom are women specifically identifiable in rock art battle scenes. Even if drawn genderless, figures in a warfare scene are assumed by most researchers to be men. However, Garcia reflected on women fighting: “There was also a hatred between the women of one tribe and the women of a different tribe. Many times a despised Indian squaw was known to stand and fight to the death by the side of her man, sometimes even against her own people” (Garcia 1967:56). In 1841, near present-day Baggs, Wyoming, in the south-central part of the state near the Colorado state line, Jim Baker observed a battle that involved a woman in a prominent position,

C. W. Lee reported Crow women acting as lookouts for possible attacking Sioux in 1868 (Hampton 2011:40), and Healy told about a Gros Ventre war party along the Missouri River between Fort Benton and Fort Musselshell (ca. 1862) where the male “Chief, followed by his squaw mounted on a war horse, was in the lead and a long distance ahead of his nearest followers” (Robison 2013:117). Thus, in rock art, women may be among the men in those genderless scenes, and it may not necessarily be the case that they are of the same tribe.

Zenas Leonard (Quaife 1978) observed warfare throughout the northwestern Plains and beyond in the 1830s and understood that it was important for people to retain their social status within the tribe. He commented on tribal competiveness that

This need to be successful for status in one’s society is behind McClintock’s observation that “the painted War Tipi of Running Rabbit was of an entirely different character, being covered with picture records of tribal victories.” He notes that “it is an interesting fact that Indians never make records of their defeats” (McClintock 1992:220–221), and today almost no rock art panels are interpreted as showing defeats. Keyser and Klassen (2001:255) provide an example from a battle scene at Writing-on-Stone that has been interpreted as the record of “Retreat up the Hill” based on 1924 information from Bird Rattle, a Piegan elder. However, there are no indications of recorded battles lost so far in Musselshell rock art.

While in the Musselshell area, Kelly spent much time staying in and describing conical wickiup war lodges. In 1869 he wrote:

He discussed another war lodge in the Bear Paw Mountains, northwest of the Musselshell and north of the Missouri River:

Yellowstone Kelly, while in the Musselshell area in 1875, wrote that “we came upon a substantial war house and concluded to camp for the night. This war house was well put up, roomy and comfortable, and had probably held twenty-five” (Quaife 1973:110). Lee also reported on pole lodges in the Musselshell area: [February 5, 1871] “up Squaw Creek about 3 miles and found where the Indians [Crow] camped . . . [They] built a lodge of dry poles . . . From the size of the lodge there could not have been over 25 Indians at the most” (Hampton 2011:1260).

Rock art representations of tipis and conical lodges occur throughout the northwestern Plains. Some rock art drawings of a single conical lodge, or occasionally multiple lodges, are made of several converging lines or many poles making up the body (figure 2.10). Such pole lodge figures do not have smoke flaps, doorways, exterior decoration, or details present on other tipi representations. Also pole lodges are usually not associated with other figures, while eagle-catching lodges are (Sundstrom 2004:124). Although structural difference may reflect individual artistic style, it is likely that isolated pole lodges shown without other interior detail or associated images depict expedient pole war lodges and not a skin-covered family residence or lodge of another function. The lack of associated context becomes the important element in functional identification of these depictions.

Figure 2.10. Conical pole lodge petroglyph at Deer Medicine Rocks (24RB401).

Long, feathered staffs in rock art are often thought to be coup sticks, and counting coup is considered a non-invasive part of warfare (e.g., Keyser 1977a, 1987a; Keyser and Cowdrey 2008; McCleary 2008a). In Musselshell rock art we identify several people shown with coup sticks. Coup sticks were often decorated with scalps (Garcia 1967:121), indicators of violence in warfare on a tool supposedly used for nonviolent contact. Yellowstone Kelly discussed a coup-counting situation in which

However, coup sticks could do damage. A Blackfoot warrior hit a Nez Perce woman on the side of her face with “his coup stick with such force as to bulge the eye from its socket, leaving it completely exposed on her cheek” (Garcia 1967:363).

Scalping was a major part of northwestern Plains warfare and is often mentioned in ethnohistorical studies. Dangling multiple lines from horse bridle bits and lances in rock art are often identified as scalps (figure 2.8). All fighters in the region, including Euroamericans, scalped their enemies. Scalping is generally thought of as being done on dead bodies only, but it occasionally occurred on someone who lived. The only Euroamerican woman living at Fort Musselshell was scalped while out with two Crow women when they were attached by a party of Sioux warriors. One of the Crow women was shot through the leg, and the white woman was shot through the neck. Thinking she was dead, the Sioux warrior scalped her, but she survived (Hampton 2011:50–51). She subsequently covered her scalped head with a wig made from red rope, suggesting another option for unusual head dresses shown in rock art.

Depictions of severed heads are not common in northern Plains rock art sites (Greer and Greer 2002). A life-sized warrior at the Daly Petroglyphs (48CA58) in northeastern Wyoming is the only one we know of on the northern Plains actually to hold a severed human head (figure 2.11). In his bent right arm he holds a bow, while his bent left arm holds the head, and he has at least one arrow entering his lower leg. The head may be held at the neck, with a feather coming out of a headdress hanging down, or the warrior is holding the top of the head by the hair with blood trickling out of the wide neck. The severed head is different from those of the warrior and other large humans next to him, presumably indicating membership of a different group. Two life-sized humans next to the warrior also hold bows and arrows, and one has a breastplate. These attributes suggest that this integrated panel portrays the results of a conflict situation.

Figure 2.11. Warrior panel at Daly Petroglyphs (48CA58). Person on far right holds a severed head. Images highlighted with Adobe Illustrator.

At least two inverted heads, seemingly severed and suspended as trophies, are at the Hewlett South site (48CK1544) in extreme northeastern Wyoming, and at least two others are at Medicine Creek Cave (48CK48), also in extreme northeastern Wyoming. These are alternatively interpreted as representing Spring Boy and Lodge Boy in Hidatsa-Crow and Kiowa mythology (discussion in Sundstrom et al. 2001:18–24, figure 11), but the heads are clearly detached.

Another possible decapitation panel is at the Manuel Lisa site (24YL82) in southeastern Montana near the mouth of the Bighorn River, where it enters the Yellowstone. Here at least five non-inverted heads are attached to a generally horizontal line by secondary cords (or perhaps weapons) to the tops of the heads (figure 2.12). Like the heads at the Daly Petroglyphs, these have distinctive hairstyles. Three have a single braid coming out of the top of the head, while two have several tassels coming out of the head. In 2002 we suggested these multiple heads may be trophies, and instead of literal decapitation may represent coup counts (Greer and Greer 2002; Keyser 2006a:62–63). Later, in 2008, McCleary, working with modern Crow and their interpretive system, suggests that the panel may be a Crow drawing of “a series of heads of enemy men and women he [the warrior on the horse above the heads] dispatched throughout his career. The first four he killed with a diamond-shaped French trade axe known as a spontoon which was favored by the Crow, and the last he speared” (McCleary 2008a:37–38). McCleary’s modern informant prefers that these are heads of dead people and not people on whom coup was counted and lived to tell about it. Although the heads at the Daly Petroglyphs and Manuel Lisa site do not have lines from the severed neck that represent blood dripping down, neither do they have attached bodies, which indicates that the person making the drawing deliberately wanted to show that the head was separated. Although we do not know if the artist was simply indicating that the people are dead, or if their heads were actually removed from their bodies at the time of death, in a battle between Crow and Blackfeet in the Big Horn Basin of western Wyoming in 1834, Zenas Leonard, who was traveling with the Crow, witnessed a decapitation associated with a battle:

Figure 2.12. Heads (previously chalked) at the Manuel Lisa Site (24YL82). The 1905 date is engraved over the image.

Conclusions

Ethnohistorical interpretations for the northwestern Plains show groups were composed of people from different tribes who would party together and then battle one another. At the time of contact many groups were using the Musselshell River drainage area, and it is highly probable that all these different groups were making rock art in the area. Ethnographies, histories written by short-term visitors, and drawings in later robe and ledger art provide information useful for image identification and inventory lists of tribes present in the area. But due to the changing cultural complexity of the region and the intensive cultural mixing, which is best described for the Musselshell country, confidence wanes when trying to link the majority of images, panels, or recurring artistic attributes with specific tribes. For instance, the seven armored horses along the Musselshell show seven different styles and are separated on the landscape, suggesting all were made by different people. Although we know the Shoshoni used horse armor as late as 1805, as observed by Lewis and Clark (Coues 1987:561), they were not the only tribe to use it (Secoy 1953). The Musselshell horses could have been put on the wall anytime from the late 1600s to the early 1800s and could have been placed there by people who owned the armor or people who observed others using it. When people of this region began drawing on robes and ledgers, it was easier to depict more detail so drawings found on these portable objects could clearly portray their tribe’s particular item of clothing (leggings, moccasins, breechcloth, or necklace) or hairstyle (hair extensions, braids, roach, etc.), but that was more difficult to do on rock and often was not included, although there are exceptions. At the Horned Headgear site on the Musselshell, a horse and rider were drawn in such detail that when Loendorf compared them with a Catlin painting he felt confident in assigning a Crow affiliation to the image, although he suggests the artist was an Assiniboine based on the detailed headdress and clothing of the person counting coup on the horse and rider, whom he believes is the artist’s self-portrait (Loendorf 2012:11–13). Thus, in order to determine tribal affiliation, style differences of costume, hair, accoutrements, and other details are needed (Keyser and Klassen 2003; Loendorf 2012). Without these clues, either because the author never included them (as such information was perceived to be artistically unnecessary), or (less likely) because details have not survived weathering of the panel, it is difficult to assign images an ethnic identity in an area where there is so much interaction and mixing of groups, as there was along the Musselshell from contact to reservation times.

Rock art evidence indicates that prior to the introduction of the bow about ad 500, weapons are scarce in northwestern Plains rock art. Before that date just as much rock art was being made (based on the number of sites recorded), but weapons portrayal was not important—atlatl figures (and even clubs) were not a topic of interest, although common in the Southwest and Great Basin. In addition, there are no identified portrayals of Archaic warfare, and the lack of intensive warfare during that time period is generally supported by skeletal evidence. After the introduction of the bow, burial sites show increased violent deaths, and weapons become prevalent in rock art. After the widespread dissemination of horses and guns in the 1700s, there is an obvious increase in warfare reflected in rock art in Wyoming and Montana, with such images as warfare scenes, interpersonal conflict, armored horses, shield-bearing warriors with weapons, and people pierced with spears and arrows. By the 1800s when traders, trappers, and the military begin recording everyday Indian life in the region, warfare was popular and familiar. However, counts of weapons and warfare images in rock art, and supported by direct observation by people living on a daily basis together with Indians (such as Andrew Garcia, Zenas Leonard, C. W. Lee, and Jim Baker), show that conflict occurred here mainly in Plains settings and in pine and juniper parklands. Warfare imagery and by extension native warfare seldom occur in mountain settings of high elevations, limestone caves, deep snow, and denser forests (Greer and Greer 2008a, 2008b). The main impetus for change was almost certainly the infusion of new groups with different practices and beliefs, and the introduction of deadlier weapons most efficient in open environments. Limestone mountains did not lose their emotional appeal as ceremonial or story-telling centers, but the sandstone-dominated plains became the main focus for rock art and its portrayal of the growing cultural importance of weapons aimed at other people. Warfare was fought by several tribes in this area to defend their territory (especially from the large groups of incoming Sioux), as revenge for killing and mutilating their fellow tribal members, to obtain goods they could not afford through trade, and for prestige, which for these groups meant elevation of status mainly through horse acquisition and coup counting. Although most battles involved small war parties of fewer than 50 people, and there was a quest for individual status, the overall view of the group being attacked was that of defending their people and preserving the honor of their tribe. Thus, the skirmishes, although small, reflected on the group as a whole, and victory benefited them all.

In conclusion, although tribal differences become harder to discern in rock art after European contact, historical documents by those living their daily life year after year with the Indians and marrying into their families provide insight into rock art interpretation different from documents produced by formal ethnographers and visitors, both of which by the 1800s viewed Indians as living museums and curiosities. Historical documents written by long-term residents come closer to providing an emic view of tribal societies making the rock art, and their reports on what was actually happening keep us from becoming too confident in assigning an explanation to a rock art panel without considering other alternatives. The diaries and narratives of area residents show that most warfare-themed rock art of Montana and Wyoming cannot be identified to a particular tribe involved in the activity because (1) there are far too many tribes in the area at the time of contact, and (2) most of the rock art dealing with warfare is too generic in how it portrays individuals, horses, and war-related activities. However, for rock art images with more detail, the descriptions and drawings by visitors and ethnographers that noted the particulars of hair, clothing, and accoutrements of the people they encountered are invaluable when attempting to make a tribal identification of these pictographs and petroglyphs. Likewise, records of early traders and trappers can offer suggestions for more broad-spectrum explanations for warfare imagery, such as that the person leading warriors into battle may not necessarily be a man when no gender is shown for the people depicted on the panel. Our consideration of whether Contact-period warfare as inferred from rock art is a reflection of warfare as portrayed in historical documents shows that these sources provide different views of conflict interactions and taken together can provide a more complete understanding of life at that time. Thus, we must continue to reevaluate our field observations of rock art panels of warfare relative to eyewitness accounts by people who lived during those times because the combined record increases our knowledge about how and why warfare was conducted in this area during that time.