It’s “Trump rebound month.” Gotta keep those Trump rebound numbers coming.

The good news here for Trumpers is that he’s cut deeply into Hillary’s lead since the last Monmouth poll a few weeks ago, a blockbuster that put him down 46/34 and sparked a thousand “Is it time for the RNC to cut Trump loose?” thinkpieces. (Monmouth is a respected pollster too, drawing an A+ rating from FiveThirtyEight, making that 12-point margin doubly ominous.) The bad news is that the new spread in today’s Monmouth poll still ain’t all that close at 46/39 in the four-way race, with Trump once again struggling to crack 40 percent. All Obama needed four years ago against Mitt Romney to engineer a national landslide was a four-point margin.

Monmouth’s national polls this summer have all had Clinton ahead, but her lead has had peaks and valleys: In June it was +7, in July (after the FBI presser on her emails) it was +2, and earlier this month amid her convention bounce it was that stunning +12 outcome. Now we’re back to +7. It looks like Hillary took a hit from Emailgate, came all the way back and then some after Philadelphia, and now has finally lost her convention bounce and settled back to the status quo ante. Can Trump cut further into that seven-point lead without external events intervening?

Clinton has the support of 85% of Democrats, a slight drop from the 92% support she had a few weeks ago. Trump gets the backing of 78% of Republicans, which is virtually unchanged from 79% earlier this month. Independents remain divided, although they now give slightly more support to Clinton (37%) than to Trump (32%). In prior Monmouth polls, Trump had the nominal advantage among independents, including a slim 32% to 30% edge in early August.

It’s surprising that Hillary’s lead is shrinking at a moment when she’s gaining among independents. It must be the downturn among Democrats that’s driving that. When Monmouth asked people whether they thought big donors to the Clinton Foundation received special treatment or were treated in an ordinary way, likely voters split 54/26, with even 26 percent of Democrats choosing “special treatment.” It may be that some progressives, having grudgingly reconciled themselves to Clinton after the convention, are feeling alienated again by the Foundation pay-for-play nonsense, a point Team Trump should note in its messaging. And in fact, when you compare the new Monmouth numbers to the last batch in early August, you find a seven-point drop in support for Clinton among self-described liberals, from 78 percent to 71 now. She’s lagging a bit among her left flank.

Here’s the most amazing number from the poll, though. Last week a pollster conducted a focus group in Wisconsin that asked people what smell they associate this election with. Among the responses he got, according to Politico, were “sulfur,” “rotten eggs,” “garbage,” “manure” and a “skunk’s fart.” After looking at this table from Monmouth, I think the correct answer is “all of the above”:

The previous record for voters who said they viewed both candidates unfavorably was set in 1992, when nine percent did. In 2016 the number is … 35 percent. Breathe deeply that sweet, sulfury skunk fart, America.

There’s something else to notice about this poll, though. When voters are given a binary choice, they prefer Clinton over Trump, 46/37, a nine-point margin. Only by introducing Gary Johnson and Jill Stein into the mix do you get Hillary’s lead down to seven. That’s another example of something I noted last week, that the third-party candidates seem to be hurting Hillary more than Trump. That’s especially true among voters age 18-34: Earlier this month, when she led Trump by 12 overall, Clinton polled 46 percent among that group compared to 26 percent for Trump and 10 percent for Johnson. A few weeks later, Hillary has lost five points among young adults while Johnson has gained seven. Trump’s total remains unchanged at 26 percent. Johnson’s presence in the race, and in particular his appeal to the under-35 group, is almost all upside for Trump right now and increasingly seems like a necessary condition for him to have any chance to win this fall. Johnson may be a better bet to poach key votes from Hillary in demographics that she badly needs to dominate than Trump himself is. If Johnson begins to fade or voters come to the conclusion that they simply must choose between the two major-party candidates, there’s every reason to believe that Hillary’s lead over Trump will begin to grow rather than shrink.

Which is to say, John Ziegler’s 100 percent right that Trump should be trying to boost Johnson right now, starting with pounding the table to have him included in the debates. That would help Trump logistically, in reducing the amount of time he’ll be expected to speak, as well as strategically, in exposing Hillary-leaning younger voters to a sustained look at the libertarian in a major TV event. Ziegler’s correct that Trump probably can’t beat Clinton head-to-head, but a three-way race where young progressives are bleeding out of the Democratic column will be less predictable. Trump has everything to gain and little to lose by raising Johnson’s profile. So why isn’t he doing it?

Update: Nice catch by Ziegler in noticing that the table above on nominee favorable ratings has called the winner of the popular vote correctly in every election dating back to 1984. If that holds, Hillary should win comfortably this year. I’d add to that by noting that, while the number who say “Democrat only” is at a modern low this year, Mondale nearly matched it in 1984 and no Democrat but Obama in 2012 was 10 points better than Hillary’s number. The “Republican only” share of 24 percent, however, is easily the worst number on the board for either party and is fully nine points worse than the next lowest GOP mark, Bob Dole’s take in 1996. Every other Republican is 12 points greater or better.