Easily the worst poll of the campaign for Trump to date. It’s a poll of likely voters, not registereds, so in theory it’s a more accurate reflection of the actual electorate than most other recent surveys. Monmouth is a highly regarded pollster, drawing an A+ rating from FiveThirtyEight. And Clinton has never before topped 46 percent in a four-way race with Trump, Gary Johnson, and Jill Stein. This survey has her blowing past that number at 50 percent, with Trump far behind at 37.

Note too, in case you think Monmouth has a persistent anti-Trump tilt, that their last poll didn’t have Clinton winning in a blowout. In July, before the conventions, they detected a two-point lead for her, which is a statistical tie. Now this. Last week I said somewhere that I thought Republican leaders wouldn’t really panic until polls start to show Trump in the mid-30s and Clinton out to a 15-point lead. Pretty close now. Clinton 50, Trump 37, Johnson 7, Stein 2:

Clinton’s 59 point lead among non-white voters is nearly identical to Barack Obama’s 58 point win with this group in 2012 (78% to 20%) according to the National Election Pool’s exit poll. However, Trump’s 5 point lead among white voters is off the 20 point mark Mitt Romney set four years ago (59% to 39%).

The main factor behind the current GOP nominee’s underperformance among white voters is his lack of support among white women with a college degree. Trump holds sizable leads among white men without a college degree (31 points; 56% to 25%), white men with a college degree (11 points; 45% to 34%), and white women without a college degree (17 points; 49% to 32%). These point spreads are similar to how Romney did with these groups in 2012 when he won white men without a college degree by 31 points (64% to 33%), white men with a college degree by 21 points (59% to 38%), and white women without a college degree by 20 points (59% to 39%). Among white women with a college degree, though, Trump is actually trailing Clinton by 30 points (27% to 57%). Romney narrowly won this group by 6 points in 2012 (52% to 46%).

That’s the whole campaign in two paragraphs. Clinton’s doing what she needs to do to replicate Obama’s winning coalition in 2012 — minorities, women, younger adults — whereas Trump’s not doing what he needs to do to replicate Romney’s losing coalition, let alone building on it to win nationally. He’s winning among white voters overall but far below Romney’s pace thanks in part to that amazing 30-point lead Hillary has among white college-educated women. The hard truth for recent Republican nominees is that they’re largely “niche” candidates, although they’re not often thought of that way because their niche, white voters, is still a majority of the electorate. Trump’s problem is that he’s a niche-within-a-niche candidate, appealing mostly to less-educated whites. (Although he does win college-educated white men here.) Either he breaks out of that niche-within-a-niche before November or somehow gets his niche to turn out at unprecedented rates or his pollsters had better be right about their theory that there are millions upon millions of secret Trump supporters out there.

In the meantime, though, expect to keep seeing results like this, at least until the debates:

Note how similar his numbers are now to what they were in March, meaning that he hasn’t done anything in five months to reassure voters that he’d be a steady hand on the wheel as president. It’s not unthinkable that, with another major unforced error like his war of words with the Khans, he’ll drop below 20 percent in that metric. To overcome doubts that deep and still have a chance to win, a candidate would need to be very, very well-liked. Trump’s unfavorable rating is 61 percent, though, the worst rating Monmouth’s seen for him yet and up eight points in just a month (again, probably a function of his battle with the Khans). The only consolation here, sort of, for Trump fans is that her lead in battleground states hasn’t really grown even though her national lead has. She led 46/39 in those states last month; now she leads 42/34. Glass half-full: Hillary at 42 percent means there are still plenty of undecideds out there in purple states for the taking.

Glass half-empty: This dude is at 34 percent, a mere one-third of the electorate, on average in states he really needs to win. This is the week Reince moves from hard liquor for breakfast to hard drugs.

If you’re inclined not to panic about any individual poll because (a) poll averages are more reliable than any one survey (smart) or (b) the polls are being SKEWED thanks to a liberal conspiracy to manipulate the data (not smart but inevitable every cycle when the GOP nominee falls behind), I’d point you to the RCP average, which now has Clinton up seven points in a four-way race. That’s easily her biggest lead of the summer. It’s also worth noting that the blockbuster McClatchy poll last week showing Clinton up 15 points seems newly credible in light of the similar margin detected by Monmouth today. It’s always more prudent to stick with the average, but there is some reason now to believe that the average might be lowballing her actual share of the vote at the moment. Either way, given the growing demand on the right for reassurance that things can’t really be this bad and that Trump is actually ahead, someone’s going to get rich before November by starting a paywalled “unskewing” site for the daily polls.

One last thing about the sample here: It’s 35D/26R/39I, which is a bit bluer than the electorate was in 2012 according to the exit polls at the time (38D/32R/29I). Doesn’t that prove that Monmouth’s poll is “biased” towards Democrats? Well, no, for the simple reason that the composition of the electorate changes every year. It may well be that there are more self-identified Democrats at the polls this fall than Republicans. In fact, given some of the discontent on the right with Trump, it may be that many people who identified as Republican in 2012 have now shifted to independent, which would explain the smaller number of Rs and higher numbers of I’s in the Monmouth numbers. Notice too that the share of Ds is also down from 2012, which might be explained by some disgusted Bernie fans shifting from Democrat to independent. It’s hard to draw firm conclusions about “skewing” from a sample unless the numbers are outlandishly out of whack with what we’d expect to see.