Interesting.
Supposing they had reached a deal, do we know how it would have worked? Would the electors vote for whichever of Fremont or Fillmore got the larger popular vote, or would the slate be divided, say 17 Fremont electors and 10 Fillmore ones. I wonder, because while I can see the Fillmore men's interest in sending the election to the HoR - Fillmore's only hope, if a remote one - I can't see what they'd get from any arrangement likely to actually elect Fremont. I'd have thought a Fremont win would scupper their party even more thoroughly than a Buchanan one.
The arrangement was as follows: There would be separate Fillmore and Fremont slates--but for 26 of the state's 27 electors the candidates for elector would be the same. The 27th elector would be either Fillmore or Fremont!--in other words, the 27th electoral vote was to be thrown away--it was just a test of who was more popular in PA. If the remaining 26 electoral votes would be sufficient to elect Fremont (or, less likely, Fillmore), then the 26 electors were to vote en masse for Fremont (or, less likely, Fillmore). If not, the electors were to divide their votes in proportion to the relative popularity in PA of Fremont and Filmore, as indicated by the vote for the 27th elector.
McClure says that if the Union ticket had been triumphant in Pennsylvania in October, the Americans would not have reneged on the agreement. He thinks it could have led to a Fremont victory nationwide, because the news that the fusion ticket had won in PA in October (1) could encourage a similar move for a fusion ticket in other doubtful northern states, and (2) might be enough to enable Fremont to carry Illinois even if there was no fusion there! This last seems a bit far-fetched to me--Fremont lost Illinois by almost four points. Still, a successful fusion ticket in Pennsylvania in November would at least send the race into the House.
Another October election was that for governor of Indiana. Ashbel P. Willard, the Democrat, won with 51.3 percent to 48.7 percent for Oliver Morton of the fusion "People's party." Note that Morton's vote was only slightly less than the combined 49.6 percent for Fremont and Fillmore in November--even though Morton had attended the Republican national convention in Pittsburgh and was by now clearly a Republican.
http://psephos.adam-carr.net/countries/u/usa/pres/1856.txt
You can imagine the sighs of relief from Buchanan supporters when the October results from Pennsylvania and Indiana became known. The Union was saved--for now! My own guess is that even fusion on the presidential level could not *quite* keep Buchanan from carrying Pennsylvania. It was one thing for a Fillmore supporter to back a Republican for Surveyor General of Pennsylvania or even Governor of Indiana. (Or for that matter Governor of Illinois, where Republican William Bissell won the governorship; he did face American as well as Democratic opposition, but the Americans did significantly less well for governor than for president.) One could do that without any risk to the Union. But at least a few of the Fillmore supporters who voted for Republicans for such state offices might have balked at a vote that potentially might help Fremont to win the presidency. (In part of course this was due to fear for the Union, but also in part to the rumors that Fremont was a Catholic.
https://groups.google.com/d/msg/soc.history.what-if/uPBeyf8ypOo/AbhEy6-T7s0J)
Now if the Democrats had nominated someone other than Buchanan and the Republicans had nominated Justice McLean rather than Fremont, the chance of a successful anti-Democratic fusion in Pennsylvania would be higher. It's no wonder that as radical a Pennsylvania Republican as Thaddeus Stevens backed McLean at the Pittsburgh convention...