The trail beyond the house
At dawn on May 7, 1880, Captain Madden’s small cavalry party and local volunteers followed a trail away from the upper San Francisco River settlement. They had come to find the people who had not found shelter. George W. Helde, a volunteer whose diary the Grant County Herald printed that month, wrote of country so rough that the riders struggled across it. By afternoon they found two Mexican sheep herders dead near the trail. Three miles farther on lay two more. Dead sheep marked the route ahead.
The Roberts house was a place where many people survived. Beyond it, the geography of the attack spread into the Mineral Creek workings, scattered grazing camps, and ranches farther up the river. A single label, “Alma Massacre,” draws those places together so tightly that later readers can mistake one family’s defense for the whole event. Madden’s May 26 report and Helde’s dated journal restore the distances. Their accounts follow men who arrived after the fighting and who had to learn what happened by following bodies, abandoned animals, tracks, and frightened survivors.
The records do not agree on every date or number. Agnes Meader Snyder remembered the attack from inside and around the Roberts house; her account was recorded in 1937. Fayette A. Jones wrote from the mining district in 1904. Madden and Helde recorded the immediate aftermath, but neither witnessed the wagon’s run to the house. Together they allow a fuller story than any one source can carry. The valley defense belongs in the first chapter. This one walks outward, toward the people whose homes and work lay beyond its walls.
The mine and its trail
James C. Cooney’s name outlived the attack. A former soldier and scout associated with Fort Bayard, he worked a silver claim in the Mineral Creek country. The trail to the mine connected workers and supplies to the valley. It also meant news of danger had to descend from the hills to households scattered around the river. A mine could have a strong cabin, a stock of tools, and armed men, but it was still a place where work finished for the day and attention shifted toward food and sleep. Agnes remembered the first attack occurring as miners quit work.
In her telling three men died at the mine. A man called Taylor was wounded in the leg and hid in a cave; others dispersed. Later Cooney and a companion called Chick came down and spoke to the Meaders. The next morning, after borrowing horses from the Roberts ranch, they went back toward the mine. The horses returned riderless. A search party met fire, and Cooney and Chick were dead. Fayette A. Jones’s 1904 New Mexico Mines and Minerals tells the sequence differently: two miners died in the initial April 28 attack; George Williams and James Taylor escaped, Taylor badly wounded; the Roberts ranch was attacked April 29; and Al Potter and John Motsinger rode out before Cooney and Chick made their fatal trip on April 30. Jones even notes that Cooney’s tomb gives April 29 while a Mr. Meeder who survived at Roberts ranch said April 30. This is an unusually useful near-contemporary disagreement. We can be confident about the mine attack, the refuge, and Cooney’s death without pretending the surviving sources settle the order of every ride.
The chronology matters because a visitor seeing Cooney’s rock tomb might imagine he died in the first rush at the mine. It matters, too, because in Agnes’s telling the borrowed horses created the second alarm. No telegraph delivered a complete account to the Roberts place. Horses covered with blood ran home without riders. The people who had loaned them could see evidence of a loss before anyone knew exactly how it happened. They sent warning to the Meaders, who still had to decide whether to shelter in their own open ground or join others below the hill. Jones’s earlier arrangement would place that ride after the main ranch attack, which means it cannot be used to confirm this exact sequence from Agnes.
Cooney’s death received a durable physical memorial. His brother and companions cut a tomb into a great boulder in the canyon, according to local history. The marker made a single death visible to later visitors. A marked grave is valuable evidence of memory and affection, but it can skew a reader’s sense of the whole episode. The herders said to have died beyond the cabin do not have equally prominent individual markers in the accounts most people encounter. The mine’s survival as a named place also creates the appearance that its owner was the central casualty. The human history was wider than the mine’s business history.
A search that kept moving
The men from Silver City left on May 4. Madden wrote that his best twenty-four soldiers rode with limited food and extra ammunition; citizens also joined the expedition. After a long ride through Mangus Springs and the Williams Valley, they reached the upper settlement about noon on May 5. Madden found roughly forty men expecting another attack. Other volunteers had gone to bring families down from Copper Creek. His report says the arrival of soldiers calmed the group. Helde’s diary supplies the human tempo behind those official words: men waited for news of bodies in the hills, prepared to ride again, and kept moving despite poor rest.
On May 6, a party recovered and buried Cooney, Chick, and Buhlman. Madden listed another man, Leasing, whose body could not be found. The burial date is firm in Helde’s diary and Madden’s report, though the exact dates of the earlier mine fights remain disputed. That contrast matters. A marker at Cooney’s Tomb later gave visitors a single day to remember. The people riding in early May were assembling a picture over several days, with each recovered body changing what they thought they knew.
At night Madden led soldiers and volunteers up the river. The local guide Foster helped them choose a route when the trail split. On May 7 the party found the four herders, then followed slaughtered sheep through rough country. The report’s precision is limited: it names no herder, and its estimate of campfires was relayed by other men. But four bodies were observed along a described path. That is stronger evidence for those deaths than a much larger total repeated without names or a route.
The search continued toward settlements on the upper river. Helde wrote on May 8 of women, girls, and herders found dead at a house; Madden’s later summary counted eight victims at the Pachico and Chavis ranch, two men, two women, and four children. Those people were not the Roberts household. Their deaths happened at another place, after or alongside the wider campaign’s movements. The reports differ in details of the tally, but both show why Alma cannot be narrated as a single siege and why a total covering several locations must be explained rather than attached to the cabin.
By May 9 the riders had reached an empty valley and then another settlement where horses had been taken. Helde’s journal turns from the dead to exhausted mounts, short rations, and families deciding whether to leave their homes. Madden’s report ends in frustration with the pursuit and the limited cooperation he believed he received from other commands. The Herald praised him and attacked General Hatch; that editorial anger tells us how exposed some territorial residents felt. It does not erase the Army’s reports of what the search party actually found.
People working sheep country
Agnes said that, after leaving the Roberts place, Apache fighters passed through country where Mexican sheepherders worked and killed thirty-five men. Her statement is central to the familiar arithmetic of the Alma episode. It is also one of the least individually documented parts of that arithmetic. The manuscript does not provide names, ages, ranches, burial sites, a list made by an investigating officer, or separate testimony from the herding families. The absence of those details in a single interview does not mean no one died. It means the most widely repeated number should be identified as a reported number until a source with a traceable roster is found.
There are nearer records, though they do not settle Agnes’s figure. A May 4 Army telegram relayed a report naming Melgoides Chavez and Jose Maria Sanchese among people killed near Old Fort Tularosa, and said Francisco Chavez had lost a son in the Mogollon Mountains along with a great number of sheep. Spellings and the exact relationship of those deaths to the Roberts-house attack need verification; the telegram gathers reports from a wide area. George W. Helde’s diary, printed in the Grant County Herald on May 22, records four dead Mexican herders found on May 7 during Madden’s pursuit. Madden’s May 26 report likewise counts four on the trail and eight additional victims at a ranch farther up the Frisco. The Herald of May 29 printed names from still wider places, including Tularosa, and a regional total for killings since May 1. Those names matter, but the May 29 list should not be relabeled as a roster of the people killed at Alma on April 28–29.
The way Agnes framed the sheepmen is troubling as well as revealing. She said they resented new settlers and had told Apache fighters that the valley’s newcomers would be easy targets. The interview does not tell us how Agnes learned of such a conversation. No named sheepman or Apache participant confirms it in the manuscript. The claim might reflect something she heard after the attack; it might be an attempt to explain why violence reached the valley; it might preserve a real dispute, but the available record does not establish that. The men allegedly killed were people with livelihoods and families, not a plot device whose deaths can be explained by an unsupported accusation. The account requires attribution rather than automatic belief or erasure. Jones’s 1904 mine history concentrates on the ranch and mining casualties and supplies no independent roster of thirty-five sheepherders.
A separate military inquiry from the wider 1880 campaign preserves a herder’s own reported words. Captain Pollock interviewed Juan Garcia after he survived a gunshot wound near the Jaralosa sheep camps. Garcia described being forced toward his camp and collapsing after he was shot; he later reached help on foot. Pollock also named Ramón Padilla, a major domo reported killed in the same broad series of attacks. These men cannot simply be added to an “Alma ranch” count—the locations and employers in Pollock’s May 9 report are different. Their names and Garcia’s survival show what a fuller record can look like when an investigator asks a worker rather than treating herders as a number. Pollock’s report is transcribed in the documentary annex, pp. 117–118.
Sheepherding in the region was mobile. Workers moved animals across grazing country, often beyond a single ranch compound. Their households could be harder for later historians to trace if a newspaper reported “Mexicans” or “sheepherders” without names. A report that dozens were killed ought to prompt a search through Spanish-language papers, parish burial registers, territorial death records, claims files, and local family histories. It ought not prompt us to treat anonymous Mexican deaths as a convenient addition that turns six into forty-one. The difference between a counted body, a reported missing person, and a number repeated from a later article is not pedantry. It determines whether the dead are being remembered as people.
The same distinction applies to the sheep themselves. A claim that great numbers of animals were scattered or killed appeared in newspaper retellings. Such losses may have devastated owners and workers, but dramatic livestock figures in war reports were often estimates made amid disorder. They cannot be used to infer human deaths. The attack’s economic consequences were real: at the Roberts ranch, Agnes remembered seventeen animals in a corral killed, and the Meaders returned to find property gone. Yet an account of loss is strongest when it stays close to what its witness could know.
A valley under pressure
The ranch house is memorable because people survived there. Its defenders mattered. Agnes described a rescue ride by six men when the Meader wagon was exposed, the dash from shed to door, people placed toward the center of the building, and a long exchange of fire. At least five Roberts family members were inside by her count. Randy Salars, grandson of Sarah Ellen Roberts Salars, preserves a family connection to the ranch. His family tree places George W. Roberts in the older family photographed near Alma, with George’s parents Josiah and Sarah E. McCready Roberts. George’s later wife Lydia Clark was only a small child in 1880 and cannot be the adult Mrs. Roberts of Agnes’s account. The five Roberts people in the house still need a named contemporary record. The ranch was a place where many residents combined shelter, water, ammunition, and watchfulness.
But staying with the house alone reverses the episode’s perspective. From the house the attackers were glimpsed as riders on a ridge, figures near a ditch, a camp in the distance. People who worked beyond sight of the building did not have its walls or its crowd of defenders. The herders Agnes described were not in the room to tell us their version. Miners killed near the initial workings were reduced to a count. A wounded worker hiding in a cave experienced a different night from the families making bullets below. A messenger riding toward Silver City experienced still another. Every route through the valley changed what a person could know and how much protection was available.
The interview gives some names worth preserving, while demanding caution about spelling. Taylor was the miner wounded in the leg. Jones identifies him as James Taylor and another survivor as George Williams. Jones names one initial dead as Buhlman and a man killed at the Roberts defense as William Wilcox. Potter rode toward the mine and was injured in the ambush; Jones calls him Al Potter and spells his companion’s name John Motsinger, where the 1937 manuscript appears as Motsiner. A defender called Foster was said to know Apache speech and signs. Murray was a man the defenders worked to bring safely into the house. Captain Madden led the relief Agnes remembered. Sheriff Whitehill later helped the Meader children travel toward Silver City. The names are not a complete roster. They are threads through documents in which many women, children, herders, miners, and Apache participants remain unnamed.
The Apache side of the country
Victorio is frequently named as the leader of the force that entered the Mineral Creek and Frisco country. To understand his campaign, begin before the mine. The Warm Springs Apache, whom the National Park Service also calls Chihennes, lived in the Southwest long before American miners and territorial officials described the land as a frontier. Their principal leaders included Mangas Coloradas, Nana, Loco, and Victorio. Their families had a home at Ojo Caliente in New Mexico. In 1877 federal authorities moved them to the San Carlos Reservation in Arizona. They left, returned, surrendered, and faced renewed pressure to move. The events drove divisions and resistance that cannot be reduced to one man’s appetite for violence.
The Fort Sill Apache Tribe explains that Victorio fought to make the United States honor agreements allowing his people to remain in their homeland. Its account also warns that a photograph often published as Victorio is likely another Apache leader named Victor. That warning has direct value here. A 1944 Desert Magazine article about Agnes Meader used a portrait labeled Victorio. Readers looking at a historical article may assume that a caption settles an identity. The tribe’s account says it does not. This series therefore does not use that portrait to illustrate Victorio or claim to recognize an individual Apache fighter at Alma from later imagery.
Warm Springs families did not move as an army detached from kin. Women and children experienced removal, pursuit, hunger, temporary refuge, and capture. The National Park Service says that after the 1880 fight at Tres Castillos, in which Victorio was killed, women and children were taken captive. The Fort Sill Apache Tribe’s portraits remember Gouyen, who was present at Tres Castillos, and Nana, who later rode to gather scattered people. Those are not footnotes to the military campaign. They are the people who give it a family history. The same is true of the Roberts, Meader, Colter, and herding households around Alma. A narrative that lets only armed men speak will miss much of what the war destroyed.
No first-person Apache account of the specific attack on the Roberts house has been identified for this series. That absence limits what can be written about the fighters’ motives, route, and decisions that day. A writer may describe the land, the federal removal policy, and the campaign’s known course; a writer may not insert a private thought into Victorio’s mind and call it history. The 1937 interview claimed that Foster understood what Victorio was telling his men during the fight. Because his interpretation survives through Agnes’s later recollection, it cannot bear the weight of a verbatim Apache command. The people whose speech was reported by opponents deserve the same evidentiary care as the settlers whose speech was remembered by descendants.
The language that came down to us
The manuscript speaks from 1937 as well as from 1880. Its language about Apache people is sometimes openly hostile. It says, for example, that Apache scouts serving the Army used government ammunition to help their relatives, and it recounts a scout later hanged by local men. Neither a blanket accusation against all scouts nor a summary execution becomes justified because it appears in an eyewitness interview. The interview does tell us what Agnes remembered and what she was willing to say to a federal writer in the 1930s. Those are historical facts about the making of the source. They are not a court finding about every individual named only by ethnicity.
The interview also contains a statement that an Apache captive told the settlers nine warriors had died. A later magazine account speaks of nine pairs of moccasins found after the attackers left. Neither statement supplies names or burial evidence. A claimed Apache casualty count should be marked as reported, just as the thirty-five herder deaths should be marked as reported. It would be inconsistent to scrutinize only the numbers that belong to one side of the story. We can say confidently that the house was attacked and defended; we cannot turn every figure recalled decades later into an audited total.
Terms shift across the records. Agnes used “Indians” as a general word and named Victorio, while today’s tribal and park histories identify particular Apache communities. The modern term Chihenne or Warm Springs Apache helps us avoid describing all Apache people as a single army. The word “Mexican” in the interview identifies the herders as Agnes understood them; it does not tell us each person’s citizenship, language, or home. “Settlers” likewise does not tell us all the prior claims to land they entered. Historical labels are evidence about a speaker’s worldview as well as possible clues about the person being described.
How forty-one entered the story
Many summaries state that forty-one people died in the Alma Massacre. The number appears in later retellings alongside Agnes’s report of thirty-five Mexican herders, but the source set does not yield a stable roster that adds to forty-one. The components are not equally well documented. Even the mine sequence varies in retellings: Agnes said three were killed during the first attack, Cooney and Chick later died, and another later account mentions a further man. A 2025 New Mexico Magazine article uses a much lower general phrase—about a dozen miners, herders, and settlers—without supplying a full roster there either. The difference is large enough that this series will not publish forty-one as a verified minimum or a precise total.
This is not an argument to minimize the violence. It is an argument to make each loss visible. A number without names may tell us an event was perceived as catastrophic; it may also carry a counting error forward for a century. A good historical account can hold a reported figure and a question together: Agnes remembered thirty-five herders killed; who were they, where were their camps, and which records could confirm their fate? The names of Cooney and Chick are not enough to close the ledger. Nor is the absence of names permission to exclude the herders from it.
What survives of the April 1880 episode has been shaped by where records were made. A federal writer interviewed an English-speaking settler survivor in 1937. A magazine writer returned to her in 1944. A regional mining history celebrated Cooney’s claim. A rock tomb kept his death legible on the landscape. Tribal histories preserve the experience of removal and exile, but do not narrate every local encounter in the form of a dated battle report. The sheepherders’ accounts remain hardest to find in this source set. The archive itself is uneven. Writing honestly about the event means showing that unevenness rather than smoothing every gap with an invented scene.
After the horses stopped running
The relief party arrived and people in the Roberts house survived, according to Agnes. The physical emergency had an end. The human consequences had no neat ending. Families needed to bury the dead, find missing workers, assess lost stock, repair buildings, rebuild tools and supplies, and decide whether to remain. The Meaders had lost property. The Roberts family had lost animals and had housed a crowd through an attack. The mine continued under Cooney’s brother; the district later expanded. Each of those outcomes belonged to a different household economy. The return of work did not mean the losses had been repaired.
For Apache families, the campaign continued through the summer. Victorio’s group fought, moved between regions, and faced forces from more than one government. In October 1880 Victorio and many followers died at Tres Castillos in Chihuahua; women and children were captured, and some survivors escaped. That later catastrophe does not answer why each individual at Alma acted as he did. It does remind us that the force passing through the Frisco country was part of a people under military and political pressure, with children and elders whose lives also stood in the path of the war.
The Roberts house survives in two kinds of image: a cabin photograph supplied by Sarah Roberts’s family and a house shaped by Agnes’s remembered words. The photograph gives us timber and fence, a building that could be repaired or razed. The manuscript gives us people moving through a ditch and a door. Neither image contains everyone affected beyond its frame. To tell the Alma story fully, the reader has to keep looking outward—to the mine, sheep ground, Apache homeland, road to Silver City, and families who carried the account into later generations.
Sources and limits
- Frances E. Totty, “The Alma Massacre,” Federal Writers’ Project, 1937, Library of Congress: the core first-person settler account, recorded fifty-seven years after the attack.
- Fayette A. Jones, New Mexico Mines and Minerals, 1904, pp. 130–132: near-contemporary named casualties, the Roberts ranch location on the west side of the Frisco, and a chronology that conflicts with the 1937 interview.
- History of New Mexico: Its Resources and People, vol. 2, 1907, p. 962: the April 28–30 span and the Cooney mining context.
- National Park Service, “Warm Springs Apache”: Chihenne removal, Victorio’s campaign, and Tres Castillos.
- Fort Sill Apache Tribe, “Our Ancestors”: a tribal account of Victorio’s goal, Nana, Gouyen, and the misidentified portrait.
- Desert Magazine, March 1944, pp. 10–12: a later, more dramatized account of Agnes’s experience that differs in details from the 1937 manuscript.
- New Mexico Magazine, “A Ghost Town with Spirit,” 2025, updated 2026: a modern comparison that uses a lower casualty description than the common forty-one.
- Robert N. Watt, Horses Worn to Mere Shadows, annex of original documents, pp. 105–110, 113–120: May 1880 Army telegrams, Helde’s diary in the Grant County Herald, Madden’s report, and the Herald’s May 29 regional casualty list. Each reports from a different date and geographic scope.
Return to the Roberts family’s house or continue to Agnes’s oral history, the family photographs, and the making of memory. The wider Victorio chapter explains the campaign beyond Alma.
Loading comments…