In North Carolina, November 5 will be almost as important as November 8. On Saturday, the 17-day window for early voting closes. Early voting in the state has so far outpaced early voting in 2012 and 2008, and the final tallies of early votes could portend a winner well before election night. The math is simple: If overall early turnout numbers increase over 2012 and 2008, and Clinton maintains a larger lead than Obama’s 56 percent in 2012 among early voters, North Carolina will be a tough state for Democrats to lose come election day.

Despite the fact that  early-voting appears to be on pace to set turnout records, there is evidence that one outcome Democrats feared—and that Republicans have engineered—might be coming to pass. After a years-long fight over voting rights, and last-minute political maneuvering by several counties, the North Carolina data group insightus reports that black turnout in the first week of early voting has been depressed relative to 2012, though it has begun to swing upward in the second week. Any slippage among this group could indicate that the historic gains in black turnout in 2008 and 2012 are in danger.

As my colleague Brentin Mock notes, there are several possible explanations for black turnout lagging at this stage in early voting, and not all of them necessarily indicate crisis. The entire state has been dealing with chaos in the wake of the Fourth Circuit Court’s decision to nullify the state’s HB 589 voting law. That decision re-established the full 17-day early-voting period from before the law, but also forced counties to figure out how to provide polling places for an extra week on short notice.

Right after most places established plans for dealing with this bit of chaos, many of the counties with the highest proportions of black voters were also hit by historic flooding from Hurricane Matthew. That storm’s impact on registration, voter awareness, poll availability, and access are reflected in coastal turnout maps, which show black voters participating in early voting at 79 percent of 2012 turnout rates. Given that geographic marginalization and disaster risk are related to the same larger marginalization that animated Jim Crow, the unique risk presented to black communities by incidents like Hurricane Matthew cannot be separated from purposeful disenfranchisement.

Mock and insightus also observe that one factor contributing to lowered black early-voting turnout could be ongoing voter suppression efforts even after the court decision on state law HB 589. The state NAACP is currently suing county boards of election over what appear to be wide racial disparities in canceled registrations, which could put black voters behind the eight ball in early voting. Also, North Carolina has 17 “rogue counties,” which managed to actually roll back polling sites or hours in the first week of early voting relative to 2012. The voter suppression theory is certainly plausible—many of those 17 counties reported hours-long lines on early-voting days—and the rogue counties include some of the largest metropolitan populations of black voters in the state. The insightus report finds that the rogue counties do have lower turnout than other counties at only 72 percent of 2012 turnout levels.

It’s important not to read too much into early-voting data before the second week’s numbers are tallied, because many counties will have more polling places and hours over this week. But the results from rogue counties so far are a kind of natural experiment in voter suppression. Even these preliminary data from rogue counties versus suggest that restricting early voting polling places and hours actually works to reduce turnout—and black turnout especially—and that expanding polling places and hours seems to increase black turnout. These conclusions seem almost like tautology, but they invalidate the tortured logic of North Carolina’s newest voting laws, which rely on the theory that curtailing the turnout-increasing mechanisms utilized most heavily by black voters would not specifically diminish black turnout.

There’s something going on beyond the laws in those 17 counties though, as the overall black turnout relative to 2012 among non-rogue counties only reached 91 percent by Monday. The most troubling theory there for voting-rights optimists is that the older, subtler means of voter suppression that seemed to be defeated in 2008 and 2012 still exist. It is still true today that voting is just harder for black voters, and always has been, even in the counties that did not establish even more obstacles by limiting polling places across the first week of early voting.

There are still generally fewer polling locations, longer wait times, and less voter information in areas where people of color live, and those same people have to work harder to arrange transportation, child care, and time off when it comes to voting at the polling sites they do have. Black voters still face various forms of voter intimidation, ballot challenges, and differentials in voter information. In North Carolina, reforms like early voting, same-day registration, and pre-registration have been intended to lower these barriers, and black organizers have found creative ways—like “Souls to the Polls” trips on Sundays—to sidestep these barriers. Yet the barriers remain, and require a good deal of organizing and individual effort to overcome.

It is certainly possible that black voters—who vote mostly Democratic—are simply less motivated to vote for Clinton than for Obama, and that record black turnout numbers from 2008 and 2012 reflect Obama’s unique appeal as a black candidate and dissatisfaction with Clinton. Obama’s candidacy tapped not into not only the sentiment of the civil-rights movement, but also into its extensive organizing infrastructure of leaders, communications, churches, and universities, a feat that would be difficult for any subsequent candidate to replicate. That historic advantage, in addition to the advent of reforms like early voting designed to boost turnout, may have established a ceiling that Clinton simply cannot reach.

But it is also possible that widespread and naked pushes to reduce black turnout since Shelby County v. Holder in 2013 have reduced faith and enthusiasm in the electoral system among some black people, even as they animate action from activist groups like the state NAACP and the “Moral Movement.” Maybe all of these things are true, and those two elections represent a high water mark, rather than the beginning of a trend.

Even after the passage of the Voting Rights Act in 1965, black voting remained depressed, and the gains of 2008 and 2012 are fragile. In those elections, turnout machines relied on well-funded and energetic organizing, and created inventive ways to use voting laws to overcome barriers for black voters. Today, activists are trying to harness anger about voter suppression and Clinton’s own historic moment in place of the magic of Obama. But turnout efforts are exhausting, and often require that poor and low-income black voters go above and beyond their ordinary election day experience. That’s no small thing to continually ask a community already beset by structural disadvantages.

Because of these factors, the 2012 black turnout numbers are probably just a bad benchmark. Read in that context, black early-voting numbers should actually inspire confidence in voting-rights activists. Even if the turnout baseline does not improve much by the end of early voting, it still points to a relatively high black turnout, historically--and one that comes in a year when voters show little enthusiasm, and in an election that lacks the once-in-a-lifetime black turnout machine that produced the historic highs.

For Democrats reliant on the strength of the black vote to open new avenues to victory and create swing states out of places like North Carolina and consolidate Florida, preliminary early-voting numbers may be cause for concern. But they should also be reminders that 2008 and 2012 were not permanent victories, and that suppression, lack of motivation, and barriers to voting have always been constant and effective pressures against black turnout. Those who rely on black votes, and those who fight for voting rights, now face the reality that much of that fight lies ahead.