In the summer of 1885, Ulysses S. Grant sat in a cottage in upstate New York, racing death to finish his memoirs. Throat cancer had been consuming him for months. He wrote in pain, often barely able to swallow.
The personal memoirs he left behind would be celebrated as one of the finest military autobiographies ever written—clear-eyed, unsparing, and frequently brilliant. Grant did not soften his failures in those pages. He did not reframe his mistakes as misfortunes or direct blame onto subordinates the way generals before and after him routinely did. But of everything he wrote across two volumes covering the bloodiest conflict in American history, one passage stands entirely apart.
Writing about the morning of June 3, 1864, at a crossroads in Virginia called Cold Harbor, Grant said this: “I have always regretted that the last assault at Cold Harbor was ever made. ”
For a man who had ordered tens of thousands of soldiers into combat across the length of the war, who had broken Confederate armies at Vicksburg and Missionary Ridge, who had understood from the beginning that winning a modern industrial war required a willingness to absorb and inflict casualties at a scale that softer men recoiled from, to single out one morning and say it should never have happened is where this story begins. Most people who know about Cold Harbor understand the outline. Grant ordered a frontal assault against fortified Confederate positions.
It failed catastrophically. Thousands of Union soldiers died in an extraordinarily short period of time, and it became one of the most debated decisions of the entire conflict. That outline is accurate. It is also insufficient.
The real depth of what Cold Harbor cost Grant—not simply in casualties, but in time, in political capital, in his army’s psychological coherence, and in the shape of everything that followed—demands a closer examination. Grant had arrived in Virginia in the spring of 1864 with a strategy that was simple in concept and brutal in what it required. He would fix the Army of Northern Virginia in place, deny it the ability to maneuver freely, and apply unrelenting pressure until the Confederacy’s capacity to sustain the war in the Eastern Theater broke. He was not indifferent to casualties, but he had grasped something his predecessors in Virginia had consistently failed to act on: the Union’s ability to replace losses was structurally greater than the Confederacy’s, and every week of sustained combat drained southern manpower at a rate the South could not sustain indefinitely.
The arithmetic, if Grant could force it, favored the Union. The Overland Campaign that began in the Virginia Wilderness in early May had already been devastating by the time the two armies reached Cold Harbor. The Wilderness had cost Grant approximately 17,000 casualties in two days. Spotsylvania had added another 18,000 across two weeks of nearly continuous fighting, including the extraordinary carnage at the Bloody Angle on May 12, a single day of combat so savage that the oak trees behind the Confederate line were reportedly cut down by the sheer volume of musket fire.
The North Anna River had added further losses. By late May and early June, Grant’s army had sustained somewhere in the range of 44,000 casualties in approximately thirty days of operations. Veteran officers who had survived three years of war were dying in quantities that outpaced replacement. Entire divisions had been reduced to the fighting strength of regiments.
The Confederate position at Cold Harbor had been engineered with a care that experienced officers on both sides could recognize immediately. Lee’s engineers had spent days constructing a network of interlocking trenches and earthworks that curved and angled across the terrain to generate overlapping fields of fire. The ground in front of the Confederate line offered limited natural cover and clear sight lines for defending artillery and infantry. What made the position particularly dangerous was precisely what made it look manageable from a distance.
It was only on closer study that the specific geometry of the earthworks became apparent, and by the time that geometry was fully understood, the assault was already committed. Grant’s plan called for a broad simultaneous attack along the front at 4:30 in the morning of June 3, with multiple Union corps hitting the Confederate line at the same moment. The intent was to stretch Lee’s ability to shift reserves. But the Confederate defensive network had been built to ensure that attacking forces would be exposed to fire not just from the front, but from angles created by the curving earthworks on either flank.
The men who were going to make that assault understood something about it the night before. Veterans who had survived three years of war had developed a particular instinct for which attacks were survivable and which were not. They had seen the terrain. They had read the ground in the way only experienced combat soldiers know how to read it.
And so in the hours before dawn on June 2 and into the 3rd, they began writing their names and home addresses on small pieces of paper and pinning them to the backs of their coats. Not letters, not final messages to the people they loved. Just names, so that whoever sorted through the dead could identify the body, so that the right family would receive the notification. When the assault began at 4:30 a.
m. , the Confederate line responded with a volume of fire that stunned men who believed they had seen everything the war had to offer. Soldiers advancing across the open ground were hit from the front and from both flanks simultaneously. Units that managed to reach the Confederate works found them impossible to breach.
Those few that briefly penetrated found themselves isolated, surrounded by fire from three sides. The best estimates place Union casualties in the first sixty to ninety minutes at somewhere between 7,000 and 8,000 men killed and wounded. The precise total is less significant than the rate. Grant called off the assault before midday.
The Confederate line was intact. Not a single objective had been achieved. Grant later wrote that when he received word that officers at the front were declining to order further charges, that men had simply refused in some instances to arise from the ground and advance again, he did not press the issue. He let the assault end.
What followed over the next thirteen days was in some respects more psychologically damaging than the assault itself. The two armies settled into static positions separated by open ground that neither side could cross without coming under fire. Between the lines lay the Union dead and wounded. Grant needed to recover those casualties, but the military protocol of the era required a formal request for a truce, and such a request carried an implicit acknowledgement that the requesting side had been unable to control the ground.
Grant was reluctant to make that concession explicit. He waited. For four days, Grant and Lee exchanged letters that circled the issue without resolving it. By the time a proper truce was negotiated and Union soldiers could move into the field, the vast majority of men who might have been saved had died.
The thirteen days at Cold Harbor did something to the Army of the Potomac that is harder to quantify than casualties but no less real. The word that moved through the ranks was not the specific language of military defeat. It was something quieter and more corrosive: the sense that the men making decisions did not fully understand what they were asking of the men carrying them out. Grant’s army did not break.
But when he finally extracted it from Cold Harbor in mid-June and crossed the James River toward Petersburg, it moved with a tentativeness that had not been there in May. The momentum of the Overland Campaign had been bled out at Cold Harbor, not in one morning, but over thirteen days of static exposure. The direct operational consequence came almost immediately. When Union forces arrived at Petersburg, the Confederate defenses were not yet fully manned.
There was a genuine window, brief but real, during which an aggressive assault might have broken through before Lee reinforced the position. The Army of the Potomac, under commanders who had just come from Cold Harbor and who were ordering men who had just come from Cold Harbor, did not seize that window with the speed and force the moment required. Petersburg held. The siege that followed lasted nine months.
Outside the army, the political damage from Cold Harbor was severe enough to threaten outcomes that extended far beyond any single campaign. The summer of 1864 was the most dangerous period of Lincoln’s presidency. The phrase “Butcher Grant” appeared in print and spread. The characterization was unfair as a full assessment of Grant’s generalship, but it found purchase because the numbers behind it were real.
Lincoln himself believed by midsummer that he would not be reelected. He wrote a memo in August 1864 acknowledging that it appeared likely his administration would fail in November. Cold Harbor had not created that crisis alone, but it was the episode that gave the peace movement its most effective single image: an assault ordered against positions that military professionals had privately concluded were not breachable, executed by men who had spent the night pinning their names to their own clothing. On the Confederate side, Cold Harbor produced precisely what southern morale most urgently needed.
Lee’s army had absorbed Grant’s campaign for an entire month and had now destroyed a Union assault in approximately an hour without yielding a foot of ground. For Confederate soldiers who needed evidence that southern independence could be won by making the North’s political will crack before the South’s military capacity did, Cold Harbor was that evidence. It was wrong in its ultimate implications, but in the summer of 1864, it was convincing enough to sustain Confederate resistance through a period when collapse might otherwise have become an internal conversation. Grant’s response to Cold Harbor in the weeks after the battle is one of the more instructive things about him as a commander.
He did not repeat it. After June 3, he never again ordered a broad frontal assault against prepared Confederate field fortifications of the kind that had failed so completely that morning. The entire operational logic of the Petersburg siege—investing the city, cutting Lee’s supply lines, extending Confederate defensive commitments until they became unsustainable—was in significant part a product of the lesson Cold Harbor had delivered about what frontal assaults against prepared positions cost relative to what they achieved. The argument has been made seriously by military historians that the Overland Campaign as a whole, Cold Harbor included, was strategically sound even where it was tactically brutal.
Grant’s sustained pressure kept Lee fixed in Virginia and established the siege conditions that eventually broke the Army of Northern Virginia. Sherman’s capture of Atlanta in September transformed the political situation. Lincoln won re-election in November, and the war ended in April 1865. The campaign, by this reading, worked.
But strategic success does not retroactively validate every tactical decision made within a successful campaign. The assault of June 3 produced nothing: no breach of the Confederate line, no strategic gain, no operational advantage of any kind. It added its thousands to a casualty total that was already testing northern political endurance to its limits, and it did so for no return that can be honestly identified. Grant understood this.
He said so. The memoirs he completed three days before he died contain the clearest possible statement of that understanding. Military leaders almost never admit in writing that a decision they made killed men for no result. Grant made that admission anyway.
He had always regretted it. He would spend whatever remained of his life having regretted it. That is what Cold Harbor was: a catastrophe, concentrated, specific, and mourned by the man who ordered it until the last day he could hold a pen.