Two weeks ago, I examined the officially meaningless but symbolically crucial popular-vote tally on the Democratic side.

At the time, Hillary Clinton had just won Pennsylvania by nine points and was claiming to be the popular-vote leader, positioning herself – in theory—to assert a moral claim on the loyalties of uncommitted superdelegates.

But the claim was based on some highly selective math—counting, for instance, Michigan (where Barack Obama wasn’t on the ballot) but not four caucus states where hundreds of thousands of Democrats participated but no official popular vote tally was kept—and I concluded that “under the most basic and probably the fairest criteria … Obama will finish the primary season hundreds of thousands of votes ahead of Clinton.”

That now looks all the more certain, not only because of Tuesday’s results from North Carolina and Indiana but also because some other widely accepted assumptions about turnout in the final Democratic contests may turn out to be wildly off-base.

In fact, as I re-compute the numbers now, Obama’s final popular-vote advantage will be, even under some worst-case assumptions for him, so large that he’ll still be ahead even if the results from Florida’s outlaw January primary are included. This is highly significant, since Clinton has justified pressing ahead with her candidacy in large part by tying herself to Florida’s (and Michigan’s) status. As some analyses have correctly pointed out, if Obama were to announce that he’s willing to abide by Florida’s January verdict, it would deflate much of the rationale for Clinton’s lingering presence.

More interestingly, Obama can also afford to count the far less legitimate results from Michigan, granting Clinton the 300,000-plus votes that she won with Obama’s name missing from the ballot—as long as he is credited with the votes that “uncommitted” received that day, a rather conservative approximation of Obama’s actual level of support among Democrats there. (Even without the uncommitted votes, Obama’s overall popular-vote lead will still probably survive Michigan’s inclusion, unless every worst-case scenario for him over the final six contests is realized.)

In short, Obama is quickly reaching the point where he can afford to make a concession on the Florida and Michigan vote that will be as inconsequential as it is magnanimous.

My original popular vote forecast—which stemmed from projections of the final eight Democratic primaries—credited Obama with a 156,000-vote combined plurality from Indiana and North Carolina. That was based on a somewhat conservative estimate of his support in both states. (I had him winning North Carolina by 12 and losing Indiana by six, but he actually won Carolina by 14 and lost Indiana by just two.) So his net popular-vote gain on the day was actually around 210,000. That’s no small difference with so few contests remaining.

The other major factor is a reconsideration of turnout in Puerto Rico, where Clinton is favored to win handily and which is roughly the same size as Oregon and Kentucky, the two largest states left on the docket. So Puerto Rico, depending on its turnout, can really skew any popular-vote forecast.

But scant attention has been paid to Puerto Rico and most analysts seem to have simply accepted the nice round estimate that 1,000,000 residents will participate in the June 1 Democratic primary.

The estimable election expert Michael Barone, in his much-discussed analysis of the popular-vote possibilities, attributed this forecast to the fact that “turnout in Puerto Rican elections is, as a percentage of those eligible, higher than anywhere on the Mainland, something on the order of 80 percent as compared with 61 percent in the 2004 presidential general election.”

But that fact might not have much bearing on this year’s primary vote. Manuel Alvarez-Rivera, an expert on Puerto Rican politics, told me that for several reasons “the estimate of one million voters for the upcoming June 1st primary is way too high.” He suggested a more reasonable turnout figure of 600,000, “give or take 100,000.”

With this in mind, I decided to run the numbers again, projecting the remaining six contests, with turnout estimates based on conversations with knowledgeable observers in each state. Here’s what I came up with:

West Virginia (May 13)<

Turnout estimate: 475,000

Rationale: The primary is open to Democrats and, for the first time, independents. But, as in California, independents must specifically request a Democratic ballot at the polls; if they don’t, they are instead handed a nonpartisan ballot that doesn’t include the presidential race. There is a gubernatorial election this year, but there is only one candidate for the G.O.P. nomination, so the Republican ballot shouldn’t be a major draw for independents. This should be a Clinton state, so let’s give her a monster victory:

Clinton: 65% (308,750)

Obama: 35% (166,250)

Kentucky (May 20)

Turnout estimate: 570,000

Rationale: Only registered Democrats are eligible to participate. Moreover, Kentucky’s absurdly early registration deadline was Dec. 31, before any primaries or caucuses were held. So while the enormous interest sparked by the Clinton-Obama race has resulted in massive registration bursts in the run-up to other primaries, there will be no freshly registered participants in Kentucky. Also, there won’t be any marquee down-ballot races to drive turnout, since Kentucky’s statewide contests were held last year. Like West Virginia, this should be Clinton country, so we’ll give her another best-case landslide:

Clinton: 65% (370,500)

Obama: 35% (199,500)

Oregon (May 20)

Turnout estimate: 655,000

Rationale: The turnout figure would be much higher, but the primary is closed, and only registered Democrats are eligible. Oregon’s elections are all conducted by mail and ballots were sent out last Friday. As of this Wednesday, nearly 90,000 had already been returned. There is also a competitive primary for the U.S. Senate on the ballot, with Jeff Merkley and Steve Novick vying to run against Republican Gordon Smith in what will be one of the most closely watched fall races. We’ll estimate turnout at 75 percent in what should be a strong Obama state:

Obama: 57% (373,350)

Clinton 43% (281,650)

Puerto Rico (June 1)

Turnout estimate: 700,000

Rationale: Manuel Alvarez-Rivera’s guess was 600,000, but since this is a Clinton state and these estimates are designed to give Clinton the benefit of the doubt, we’ll go with his high-end projection of 700,000. We’ll also add this caveat, courtesy of Alvarez-Rivera: “I tend to agree with the conventional wisdom that Sen. Clinton should prevail easily in Puerto Rico, but I don't rule out an Obama victory: sometimes voters deliver a totally unexpected outcome, as they did in 1988, when Jesse Jackson won the Puerto Rico presidential primary to everyone's surprise.”

Clinton: 60% (420,000)

Obama: 40% (280,000)

Montana (June 1)

Turnout estimate: 210,000

Rationale: Turnout was 37 percent the last time around, so we’ll bump it up to 50 percent for what will be the state’s first consequential presidential primary in decades. There is no party registration in Montana, but the state obviously tilts toward the G.O.P. However, the same can be said for Indiana, where about 75 percent of all primary participants opted for the Democratic ballot, partly because there was no action on the Republican side. There isn’t much more going on for the G.O.P. in Montana, but we’ll estimate that only 65 percent of primary voters take a Democratic ballot. Like every state around it, Montana should be a lay-up state for Obama:

Obama 61% (128,100)

Clinton 39% (81,900)

South Dakota

Turnout estimate: 90,000

Rationale: Why the big drop in turnout from Montana? Because South Dakota is a closed primary state. There are about 200,000 registered Democrats in the state, and we’ll go with 45 percent turnout (again, trying to err on Clinton’s side here, since this should be Obama country).

Obama: 61% (54,900)

Clinton: 39% (35,100)

So where does all of this lead us? Working off of these projections, Clinton will win about 295,000 more votes than Obama in the remaining primaries—just about half of them from Puerto Rico. And keep in mind, this is all based on Clinton-friendly projections. If Obama can hold her to 60 percent in West Virginia and Kentucky, he’ll shave more than 100,000 votes from her edge. And just three extra points in Oregon could trim another 50,000. And don’t forget that the turnout estimate for Puerto Rico may be high, to say nothing of Clinton’s projected margin there.

But let’s stick with the original projections, with Clinton gaining nearly 300,000 votes between now and June 3. What does this do to her effort to claim a popular-vote victory?

Under the most basic, logical and fair popular criteria—only counting the states and U.S. possessions that held official contests in which exact popular-vote totals were maintained—Obama’s lead would be 414,278 votes. In other words, his lead under this criteria is now iron-clad, even based on projections based on what should be worst-case data.

Now let’s look at Florida, Michigan, and the four states (Washington, Maine, Iowa and Nevada) where caucuses were held but no official popular-vote tally was kept.

If we just count Florida, Obama’s final lead would shrink to just under 120,000 votes.

But if we’re going to count Florida, there’s really no justification for not counting those four caucus states, where generally accepted popular-vote estimates have Obama generating a plurality of 110,222 votes. Factor them in and Obama’s final advantage—with Florida—almost doubles.

Now let’s look at Michigan, a deeply flawed contest where turnout was exceedingly low and Obama’s name was absent from the ballot. Clinton “won” the primary there with 55 percent, while “uncommitted” netted 40 percent. Polls, though, have shown that Obama would have run even with, and very possibly ahead of, Clinton in a legitimate primary.

Still, for the sake of argument, let’s award Clinton the 328,309 votes she claimed in Michigan back in January and give Obama zero. When you factor that in with Florida and the four caucus states, Clinton would take a slim 98,000-vote lead in the overall popular vote. But, again, this is based on worst-case data for Obama. If he holds Clinton to 40 percent wins in West Virginia and Kentucky, he’ll end up ahead in the popular vote even when only Clinton’s Michigan votes are counted (albeit by just a few thousand votes).

Of course, only counting her votes is unfair. A reasonable compromise for including Michigan would involve giving Obama credit for the uncommitted votes, many of which were intended for him anyway. Again, since polls have shown him running even with Clinton in the state, this is a conservative expression of his support. But if we give Obama the “uncommitted” 238,168 votes, he pulls ahead in the overall popular vote by almost 140,000 votes.

So even if we assume the worst about how Obama will do over the next few weeks, Obama will still lead the popular vote even in Florida is counted. And he’ll lead it with Michigan too, as long as he’s credited with the uncommitted votes there. And even if he’s not, he’ll still probably win the popular vote, if he outperforms my meager forecasts for his totals in the remaining primaries.

After Pennsylvania, the Clinton campaign was quick to talk about her supposed popular-vote lead. It’s not an accident that they’ve been so quiet on the subject this week.