On September 13, 1862, two Union soldiers resting in a field outside Frederick, Maryland, found something that should have ended a campaign before it truly began. Corporal Barton W. Mitchell and First Sergeant John M. Bloss of the 27th Indiana Infantry discovered an envelope wrapped around three cigars.
Inside was a copy of Robert E. Lee’s Special Orders No. 191, the complete battle plan for the Confederate Army of Northern Virginia, then split into four columns spread across nearly sixty miles of Maryland countryside. The document reached Major General George B.
McClellan’s headquarters the same day. Within two days, McClellan had used it to fight his way through the mountain passes at South Mountain. Within four, he had brought Lee to bay along Antietam Creek. What McClellan did with that intelligence defined the rest of the campaign.
He had Lee’s order of battle in hand, yet his estimate of Confederate strength, supplied by his own private intelligence service under the detective Allan Pinkerton, put Lee’s army at nearly 100,000 men. The true figure was somewhere between 38,000 and 45,000. The number guiding every decision McClellan made that week overstated the enemy by roughly two and a half times. And that figure went unchecked, unverified against the ground in front of him, as it had been before and would be again.
McClellan had risen on the strength of that habit. Born in Philadelphia in 1826, he had graduated second in his West Point class, served on Winfield Scott’s staff in Mexico, observed the Crimean War, and run the Illinois Central Railroad before the Civil War brought him back to uniform. In the summer of 1861, after the Union rout at Bull Run, he took command of the shattered volunteers gathering around Washington and turned them into a disciplined, confident army. That achievement was real.
It convinced Abraham Lincoln to place him in command of the Army of the Potomac and then, in November 1861, to name him general-in-chief of the entire United States Army, the youngest man ever to hold the post. But the instincts that served McClellan in organizing an army failed him whenever he faced an enemy he could not precisely measure. The pattern surfaced first in western Virginia in July 1861. At Rich Mountain, he judged the Confederate force before him at up to 50,000 men.
The detachment actually holding the pass numbered roughly 4,000 to 5,000. The victory that followed came from Brigadier General William Rosecrans, who flanked the position while McClellan held his main force in front. The record recorded it as a McClellan victory. No one asked how a fivefold miscalculation had shaped the plan.
The same failure reappeared on a larger stage outside Yorktown in the spring of 1862. Confederate General John Magruder held the town with about 11,000 men but made that garrison look far larger by marching troops back and forth in view of Union lines and mounting logs painted black to resemble heavy guns. Pinkerton’s agents reported what those Quaker guns implied. Balloon observers under Professor Thaddeus Lowe added estimates pointing the same direction.
Every source McClellan consulted agreed because every source had been shaped by the same unexamined assumption. By late August 1862, the pattern extended even to armies McClellan was supposed to be helping. On August 29, he telegraphed Lincoln proposing that rather than rush reinforcements to General John Pope’s endangered army, the government might “leave Pope to get out of his own scrape” and concentrate instead on the safety of Washington. Nineteen days earlier, in a private letter to his wife Ellen, he had predicted Pope would soon be badly beaten and that Washington would then be glad to hand the war back to him.
The Army of the Potomac that fought at Antietam was McClellan’s creation, rebuilt from the chaos of Bull Run. Its defeat at Second Bull Run had left Washington exposed, and Lincoln had turned to McClellan again. What followed should have been the campaign’s decisive moment. After South Mountain on September 14, the two armies met in full on September 17 along Antietam Creek near Sharpsburg.
The fighting moved across ground fixed in memory by name: the Cornfield, the Sunken Road, the Stone Bridge. By nightfall, Lee’s army remained where it had stood when the fighting began. The armies had lost 22,717 men killed, wounded, or missing in a single day, the bloodiest day in American military history. Union losses came to 12,410; Confederate losses to 10,316.
Six general officers were gone. Among the wounded was Corporal Mitchell, the man who had found Lee’s orders. His leg wound never healed cleanly; chronic infection forced his discharge from the army in 1864. He died in 1868 at fifty-two.
His widow did not receive a pension until 1890. The next day, September 18, was the decision that defined the campaign. Across the field, Lee’s army stood in place, braced for a renewal of the assault that never came. Union Major General William Franklin’s Sixth Corps, nearly 20,000 men, sat in reserve untouched by the previous day’s fighting and received no order to advance.
McClellan had Lee’s orders, four days out of date, and estimates of Confederate strength nearly two and a half times the real number. He had no reliable count of what fresh troops, if any, remained behind Lee’s lines. That gap he showed no impulse to fill. Only under cover of darkness that night did Lee’s army begin its organized crossing back into Virginia at Boteler’s Ford.
The next day, McClellan sent cavalry forward under Brigadier General Alfred Pleasonton with written instructions not to cross the river “unless a clearly favorable opportunity” presented itself to strike without loss. It was, in its own language, a standing order against risk. The caution did not prevent the fighting from resuming. On September 19 and 20, Union troops crossed at Boteler’s Ford, and A.
P. Hill’s Confederate division turned back to meet them. The heaviest cost fell on the 118th Pennsylvania, the Corn Exchange Regiment, which lost well over a third of its strength in the engagement. On November 5, the day after the midterm elections, Lincoln signed the order relieving McClellan of command.
Two days later, in the middle of a snowstorm, Brigadier General Catharinus Buckingham rode into McClellan’s headquarters at Rectortown, Virginia, and placed the order in his hands. McClellan’s own description of the moment, written that evening to his wife, insisted not a single muscle in his face had moved as he read it. He admitted he had made mistakes but could not identify any he considered serious. Three days later, the army gathered at Warrington for a farewell review.
A column said to have stretched some three miles filed past him, and by several later accounts, many of the men in the ranks were in tears. McClellan never returned to field command. In 1864, he accepted the Democratic nomination for president and lost to the man who had removed him. What explains the decision at Antietam is not one failure but a structure.
The strongest explanation is institutional. Pinkerton’s intelligence operation functioned in practice as a business with one safe direction to be wrong. A report understating Lee’s strength, later proven wrong, could cost lives and careers. A report overstating it cost nothing; it simply confirmed the caution McClellan already leaned toward.
No incentive pushed toward a lower number, and nothing in the surviving record suggests one ever did. The second explanation is psychological. McClellan’s letters to his wife return again and again to a conviction that he alone understood the true situation and that Washington did not. A man convinced of that isolation has little reason to interrogate a figure that supports what he already believes.
The third explanation, genuine limitations of military intelligence in 1862, cannot account for direction. An honest instrument limited by its technology should have been wrong in both directions. McClellan’s numbers ran high consistently for an entire war. That consistency is clearest in the case of Malvern Hill.
On July 1, 1862, during the Seven Days Battles, Lee’s army attacked a fortified Union position and was thrown back with heavy loss. Confederate casualties came to 5,650 against 3,007 Union. It was a clear tactical victory for McClellan’s army. But for most of the fighting, McClellan was not on the ridge.
He had gone downriver aboard the ironclad USS Galena, inspecting a new supply base, while his Fifth Corps held the line above him. Fitz John Porter, the man commanding that line, argued directly to McClellan that one more push could carry them to Richmond. McClellan declined. He ordered the army to continue its withdrawal toward Harrison’s Landing, away from the capital his own field commander was urging him to take.
The reasoning matched point for point what would return two and a half months later outside Sharpsburg: the conviction, resting on Pinkerton’s figures, that the army in front of him remained larger than his own. At Antietam, McClellan stood near the field and chose caution with the fighting still visible around him. At Malvern Hill, he made the decision entirely in the abstract, built on a number he trusted from a distance, about a battle he did not personally watch his own army win. Lincoln saw the Army of the Potomac for himself in early October, walking the ground at Sharpsburg and meeting with the general who had commanded there.
A remark from that visit has survived in more than one retelling: standing over the rows of tents, asked what he was seeing, Lincoln called it not the Army of the Potomac but something closer to McClellan’s own guard. The wording varies by account; the absence of warmth is consistent. A more solidly documented exchange followed late that month when McClellan cited exhausted horses as a reason the army could not move. Lincoln’s telegram asked, in effect, what the horses had been doing since Antietam to account for such fatigue.
Two things belong on McClellan’s side of the ledger. He built from nothing the first professional standing secret service the United States Army had fielded, a disciplined intelligence operation in place of the rumor that had passed for reconnaissance before him. And nothing in the surviving record suggests he ever doubted the figures that operation produced. He commissioned them, paid for them, and believed them completely.
Deliberate sabotage was not what failed. What failed was structural: an intelligence service never held accountable for overstating a threat, and certainly held accountable for understating one, will eventually be believed by a commander with every reason to want it true. At that point, caution stops being a trait of the man giving the orders.
It becomes the output of a system he built so that he would never have to check its work.