The last time Clinton had two consecutive polls in Pennsylvania putting her lead at nine points or better was in early August, during her convention bounce. The fact that two respected pollsters in Monmouth and Franklin & Marshall have her climbing back to that level now, a month out from Election Day, is grim news for Trump. Worse, Monmouth has her hitting 50 percent in the four-way race, something she’s never done before in Pennsylvania. Franklin & Marshall has her at 47 percent, also among her best numbers in the state in 2016.

If Pennsylvania turns into a 10-point race consistently in polling over the next few weeks, what’s the play nationally for Trump? She can win the election by winning this state, the usual blue strongholds, plus Virginia, Colorado, and New Hampshire. Trump hasn’t led a poll of New Hampshire all year. The same is true of Virginia, where he hasn’t done better than a six-point deficit in recent weeks. He has led in Colorado recently — but not since the debate. In fact, in the only two polls of that state since last Monday (one of them conducted by a Democratic firm), Clinton’s led by 11 points. One of these states needs to change, dramatically and soon, or else Trump is stuck trying to peel off a traditional blue stronghold. Er, which one?

Same story in both polls, by the way. On basic qualifications for the office, Clinton’s lead over Trump is vast. Monmouth:

Among Keystone State voters likely to cast ballots in November’s presidential election, 50% currently support Clinton and 40% back Trump. Another 5% intend to vote for Libertarian Gary Johnson, 2% back Jill Stein of the Green Party, and 2% are undecided. In late August, Clinton led Trump by 48% to 40%…

Clinton has the edge over Trump on being seen as someone who understands the concerns of typical voters. Most Pennsylvania voters say Clinton has this characteristic – 52% say she does and 46% say she doesn’t. Fewer see Trump as someone who understands the day to day concerns of people like them – 38% say he does and 60% say he doesn’t…

Clinton has an even wider advantage on the issue of presidential temperament – something that has been dominating the news since last week’s debate. Twice as many voters say Clinton has the right temperament for the job (64%) compared to those who say the same about Trump (31%).

The results on presidential temperament are nearly mirror images, with Clinton at 64/33 and Trump at 31/66. The big shift overall in the state since August is among white women, who used to split 46/45 for Clinton and now break 55/35 for her, which may be a cause or an effect (or both) of Hillary’s Alicia Machado gambit. Republicans may scoff at Clinton’s gender panders in attacking Trump but there’s no reason to doubt that they’re working. In fact, I’d be curious to know how many Republican women specifically have gotten cold feet about Trump over the last month and a half. In August, Monmouth found GOPers breaking 81/9 between their own nominee and Clinton. Now the split looks like this:

Sixteen percent is a big number when it comes to members of one party crossing over for the other. I’d bet it’s white suburban Republican women first and foremost who have decided to cut Trump loose. Those are the voters he was courting with his outreach to minorities and his “softening” on immigration in August. It looks now like that effort’s been undone. Note too that Gary Johnson is at just five percent among Republicans, less than a third of what Hillary is pulling. That suggests more than just casual disaffection with Trump among anti-Trump GOPers, since they could park their votes with Johnson if all they wanted to do was not vote for their own nominee. Crossing over for Clinton is a sign that a small but meaningful number of Rs is actively trying to beat him. He’d have to have some debate on Sunday to turn that around.

As for Franklin & Marshall, they have the race 47/38 for Clinton now. In August it was 45/40.

Note the last item at the bottom. Check the crosstabs at the link and you’ll find that Trump has made virtually no progress on most of the “presidential attribute” questions since July. He showed some modest improvement in August, but now he’s slid back to where he was three months ago. The trends are going in the wrong direction as time ticks down. Also of note: Monmouth and Franklin & Marshall both have Clinton ahead now in Pennsylvania among white voters, Trump’s base, thanks to her strength among white women. Monmouth has her ahead by a single point among whites whereas F&M has her up by eight points(!!). The latter poll has him winning whites without a college degree, but by a comparatively slim margin of seven points. White college grads, by contrast, break hard for Clinton at 54/32. Again, it’s better-educated white suburbanites who have turned on Trump.

One more number, this one a measurement of leaners by Franklin & Marshall. Compare Gary Johnson’s trendline to Trump’s:

You don’t want your guy finishing third among undecideds when his best-case scenario is a very narrow win. But that’s the state of the state in a nutshell. Clinton’s favorable rating in Pennsylvania right now in both polls is close to breaking even — 41/48 in Monmouth, 47/50 in F&M — but Trump’s is a horrible 27/60 and 32/60, respectively. If he can turn all of that around in a month and win the state, it’d be one of the great political comebacks of all time. He has 35 days.