In the aftermath of the Battle of Gettysburg, the Confederate Army was in full retreat, forced to abandon all of its dead and most of its wounded. The Union Army and citizens of Gettysburg had an ugly cleanup task ahead of them. Along with the numerous corpses littered about the battlefield, at least 27,574 rifles (I've also seen 37,574 listed) were recovered. Of the recovered weapons, a staggering 24,000 were found to be loaded, either 87% or 63%, depending on which number you accept for the total number of rifles. Of the loaded rifles, 12,000 were loaded more than once and half of these (6,000 total) had been loaded between three and ten times. One poor guy had reloaded his weapon twenty-three times without firing a single shot. At first glance, this doesn't seem to make any sense whatsoever.
One could draw any number of conclusions from this data. But an obvious one might be that for some reason, large numbers of soldiers were not discharging their weapons during the battle but continued to reload anyway, perhaps to give off the appearance that they were participating in volleys. The thick smoke, mass confusion, and thunderous sounds of musket and cannon fire would probably prevent a neighbor on your line from definitively observing that you weren't actually firing your weapon. You could even mimic the rifle's kickback as you pretended to fire. In his book on the psychological impact of killing in war, On Killing, Dave Grossman argues this very point, coming to the conclusion that the discarded but loaded weapons recovered after Gettysburg mostly represent soldiers who were psychologically unable or unwilling to fire at the enemy.
Paddy Griffin highlights a few other possibilities in her well regarded book, Battle Tactics of the Civil War. For one, Griffin argues that the high rate of misfire in Civil War era rifles combined with the inability of many soldiers to reload properly under hectic battle conditions would render a large number of rifles unusable in a short period of time. Loading a civil war rifle, such as the Springfield 1861, was a complex and time consuming process. In the heat of battle, it is to be expected that some number of soldiers will panic, lose focus, or act in fear, leading them to misload and thus misfire their weapon, rendering them useless.
The rifles of the era were prone to overheating and often malfunctioned on their own and the rate of misfire only increased with each successful shot. Inserting the percussion cap, the final step before firing, was easy to bungle or forget, potentially leading a soldier to think that he had discharged his weapon when he hadn't. There is also the chance that a soldier accidentally fires his ramrod (essential for reloading), then begins to reload his weapon only to find he cannot complete the job. These weapons would likely be abandoned. A new weapon would be claimed but it too could be a discard. The soldier would reload the newly acquired weapon only to find that it cannot fire, and then immediately drop it. Now the rifle is double loaded. There is anecdotal evidence to suggest that in some instances the misfire rate could be as high as 25% during combat. We might believe then that many of the loaded rifles were discarded on the battlefield precisely because they couldn't fire. Functioning rifles might be desirable enough to be claimed from the battlefield long before an official tally of leftover weapons was made.
Finally we must consider the high attrition rate of artillery fire, which could engage the enemy at much longer ranges than musket volleys. During Pickett's Charge, the confederate army marched slowly towards the union lines and only began anything resembling a spirited jog once they had closed to a few hundred yards. Throughout the war, both sides were reluctant to fire until they had their intended target within their sights. By the time they reached volley range, cannon fire would have already decimated whole sections of the line, leaving behind dead or dying men clutching fully loaded rifles.
These factors probably all contributed. It's certainly believable in light of other studies that some percentage of soldiers intentionally fired over the head of the enemy, or perhaps double, triple, or quadruple loaded their rifles to avoid firing them at all. But 90 or even 60%? That seems ludicrous. The number of casualties at the battle alone (33,000 between the two sides), not all of which could have been caused by artillery, attests otherwise. Those who misloaded or misfired their weapons were among the lucky ones. Plenty were killed before they could fire off a single pre-loaded shot.