Bernie Sanders answered two important questions with his strong showing in Iowa. But, despite his impressive finish, he’ll need to answer two more to truly threaten Hillary Clinton for the Democratic presidential nomination.

The most powerful lesson from the Iowa caucus results is that Democrats are facing not just a generation gap, but a Grand Canyon-sized chasm. As I wrote this week, age has emerged as the single most important dividing line in the struggle between Sanders and Clinton.

In the Iowa entrance poll (which questions voters on the way into a caucus, rather than on their way out the door, like “exit polls” in primaries) Sanders amassed astounding margins among young people. He crushed Clinton by an almost unimaginable six to one—84 percent to 14 percent—among voters younger than 30. For those tempted to dismiss that as just a campus craze, he also routed her by 58 percent to 37 percent among those aged 30 to 44.

But Clinton’s margins were almost as impressive among older voters: she beat Sanders 58 percent to 35 percent among those aged 45-64, and by 69 percent to 26 percent among seniors.

That’s an even wider age gap than Iowa produced in the 2008 contest between Clinton and Barack Obama. In that Iowa caucus, Clinton also was routed among younger voters, but Obama stayed more competitive than Sanders did among those older than 45. On both sides, John Edwards, as a strong third contender, also somewhat muted the contrasts. In 2008, Clinton ran 34 percentage points better among seniors than with those under 30; this week, the gap was 55 points.

Obama beat Clinton by 20 percentage points among voters younger than 30, while she beat him by 25 points among voters older than 65, according to a cumulative analysis of the results of all the exit polls in the 2008 Democratic primary conducted by ABC pollster Gary Langer. Voters in the middle-aged groups divided more narrowly: Obama carried those aged 30-44 by 11 points, and Clinton carried the near retirement generation (45 to 64) by seven, according to Langer’s analysis.

But when it comes to piling up votes, one of these demographic advantages is much more useful than the other. Across all of the 2008 contests, according to Langer’s calculations, voters older than 45 cast fully 61 percent of Democratic votes, while those younger than 45 cast 39 percent. That’s an advantage for Clinton. And it’s a slightly worrisome note for Sanders—a cloud passing on an otherwise sunny day—that young voters cast a slightly smaller share of the total Iowa Democratic vote in 2016 than 2008.

Still, Sanders’s overwhelming margins among Iowa’s younger voters—which exceeded even Obama’s 2008 showing—affirmatively answered the first critical question for the Vermont senator’s campaign: Would the connection with young voters evident at his rallies translate to the ballot box?

The second question that Iowa answered is whether Sanders could become more than just what I’ve termed a “wine track” candidate. Sanders started the 2016 race fitting the profile of brainy Democratic contenders from Eugene McCarthy to Gary Hart to Paul Tsongas to Bill Bradley, who relied mostly on support from younger voters and upscale white liberals. All of them lost to rivals who mobilized the competing “beer track” coalition of blue-collar whites and minorities. Obama represented a variation on the theme: He married the backing of those upscale whites with support from African Americans. That allowed him to narrowly beat Clinton’s modified “beer track” coalition of working-class whites and Hispanics.

In Iowa, Sanders did very well with the wine-track constituencies. Besides his overwhelming showing among younger voters, he stayed close to Clinton among college-educated whites: He carried college-educated white men by 12 percentage points and lost college-educated white women by the same margin. In Iowa, in fact, both college-educated white men and women divided almost exactly as they did in the cumulative results between Clinton and Obama in 2008.

But in Iowa, Sanders expanded beyond that beachhead to run evenly with Clinton among non-college whites. If that pattern persists, it would represent a huge change in the Democratic landscape. In 2008, across all the exit polls, Clinton dominated Obama among non-college white women (carrying 66 percent of them) and non-college white men (56 percent). But in Iowa, Sanders narrowly carried those blue-collar white men and held down his losses among the blue-collar white women: Clinton only carried just over half of them.

Class and age shaped the gender gap in Iowa. That gap was modest: Clinton last night won 50 percent of women and 44 percent of men. That almost exactly matched her total in the cumulative 2008 results against Obama, when gender was also somewhat muted: In that race Clinton carried 52 percent of women and 43 percent of men.

Clinton benefits from that pattern because women cast most of the Democratic votes (57 percent in all the primaries last time, the same number as in Iowa last night). And her overall advantage among women could rise as African American women, who usually vote more heavily than black men, weigh in. But in Iowa, Sanders signaled he could remain competitive enough among women (particularly younger woman) to prevent Clinton’s advantage there from becoming insurmountable.

It’s yet to be seen whether Clinton’s performance with blue-collar whites eroded because those voters are responding to Sanders’s full-throated economic populism—or simply because she’s not running again against an African American with an Ivy League pedigree. But, either way, if Sanders can sustain the competitive showing among blue-collar whites he displayed in Iowa, he can contest metal-bending Rust Belt states (like Ohio, Michigan, and Wisconsin) that previous wine-track candidates could only rarely threaten.

But to win those states—and the other big racially-diverse states that will almost certainly decide the Democratic race—Sanders must answer two other questions Iowa did not resolve.

The first is whether he can win minority voters. Minorities comprised only 9 percent of the voters in Iowa last night (up just slightly from 2008), but they will likely cast between 35 to 40 percent of the total vote in the 2016 Democratic primaries. As I noted on Monday, minorities are especially plentiful in the big states that will award the most delegates, including New York (where minorities cast just under one-third of the 2008 vote), Florida (about one-third), Virginia and New Jersey (about two-fifths), Illinois (over two-fifths) California (nearly half) and Georgia and Texas (over half).

The sample in Iowa was small, but Sanders won only about one-third of non-white voters there, compared to about three-fifths for Clinton. She polls even better among minorities in most national surveys. The next contest in New Hampshire, whose Democratic electorate in 2008 was 95 percent white, won’t provide much guidance on whether Sanders can shatter that wall. The real signals will come later in February from Nevada (where Hispanics and blacks each cast about one-sixth of the 2008 vote) and South Carolina (where African Americans cast a 55 percent majority of the 2008 vote).

Because of Clinton’s continued strength with white women, Sanders almost certainly can’t amass margins large enough among all whites to win big states if he can’t make further gains among minorities. Sanders’s campaign sees more opportunities with Hispanics than African Americans. But from whatever camp they’re drawn, winning more minorities in the big states looming on the calendar is the first key test of whether Sanders can truly threaten Clinton.

The second is whether he can win more Democrats. Sanders won over two-thirds of independents who participated in the Iowa caucus. But even amid his otherwise strong performance, he lost Democrats by a resounding 56 percent to 39 percent. Compared to Obama in Iowa in 2008, Sanders enjoyed a wider margin among independents, but fared much more poorly among Democrats: Obama and Clinton split them about evenly eight years ago.

This profile won’t hurt Sanders in New Hampshire, where independents (and Republicans) cast nearly half of all the votes in the Democratic primary last time. But it’s hard to win a party’s nomination without performing competitively among voters in that party. Many states restrict participation to registered Democrats. In 2008, self-identified Democrats cast almost exactly three-fourths of Democratic primary votes, and Obama held Clinton to a narrow 6-percentage-point advantage among them—allowing him to make up the difference with his crossover support.  Sanders, who did not describe himself as a Democrat until recently, is very unlikely to become the Democratic nominee without converting more Democrats to his “political revolution.”