Oops, I forgot about that line you mentioned!  Anyway I think White does have an improvement, 9.Qe1 (instead of 9.h3 Bxf3+ =) 9...Nxf3 10.Qxh4+ Nxh4+ 11.Ke1 f3 12.d3 (or 12.h3, but not 12.d4 fxg2 13.Bxg2 Nxg2+ 14.Kf2 Nh4!) 12...fxg2 13.Bxg2 Nxg2+ 14.Kf2 Kc8 15.Kxg2 Kc8 and White comes out with rook and pawn for two pieces (a rough material balance) and has better development and central control, so probably has some advantage.

Your ideas in the Hampe-Allgaier are interesting- with 9.Bc4+, 11.0-0 and 11.Qd2 all looking playable, it all goes to show that the opening is in a decent theoretical state- certainly better in my opinion than the positions resulting from 3.Nc3 Qh4+ 4.Ke2 Qe7 (has anyone got any ideas for White there?)

Perhaps White should try 11.0-0 instead of 11.Qd2 (after which I can't improve on your analysis or assessment).  For instance, 11...Bxc3 12.bxc3 Kg7 13.Qd2 and White has a strong centre and pressure down the f-file to compensate for the material deficit.  Or 11...Kg7 12.Ne2, planning Ng3 and also c3 chasing away the a5-bishop, e.g. 12...Nh5 13.c3 Ba5 (not 13...Nxf4?! 14.Nxf4) and now perhaps 14.Qb3 or 14.Bc4.

It's possible that Black could still be better with best play here (Fritz 10 assesses the resulting positions as 0.6 to 0.8 of a pawn better for Black), but play is more open-ended, it certainly isn't an easy route to advantage for Black. 

Fritz 10's first choice isn't 10...d6 (at least not initially, perhaps if it is run for a long period of time it might change its mind?) but it does look like a more critical continuation than those given in the previous thread.

Meanwhile, although I'm not completely familiar with the theory after 3.Nc3 Qh4+ 4.Ke2, my impression is that the main line 4...d5 5.Nxd5 Bg4+ 6.Nf3 Bd6 7.d4 Nc6 8.c3 0-0-0 9.Kd3 Qh6 10.Kc2 leads to outright chaos, where White's strong centre compensates for the lost pawn, but that 4...Qe7!?, intending ...Nf6 and ...d5, may give Black an edge.  Fritz 10's openings book gives 5.d4 Nf6 6.Kf2 (better is 6.Bxf4 Nxe4 7.Kf3(!) Nxc3 8.bxc3 d5 9.Kf2, but this still favours Black) 6...d5 7.e5?! Ng4+ which is excellent for Black.  5.Nf3 is an alternative for White, but if for example 5...d5 6.Kf2 dxe4 7.Nxe4 Nf6 8.Nxf6+ Qxf6 9.Bb5+ c6 10.Re1+ Be7 11.Bc4 0-0 I think Black still stands better.

There was an analysis of the Hampe-Allgaier at this thread:
http://www.chesspub.com/cgi-bin/yabb2/YaBB.pl?num=1217111507/19#19   10.Bd3 is a better option for White in your first line, and that the H-A-G is currently looking unclear/equal.

White also has the Pierce Gambit (5.d4) but after finding improvements for Black over Tim Harding's analysis I think it's unsound.

However I agree that 3...Nc6 is a good alternative to the 3...Qh4+ lines.

I thought Black had a forced draw:
Welling,G - Van Mil,J [C33]
Eindhoven, 1973
1.e4 e5 2.f4 exf4 3.Nc3 Qh4+ 4.Ke2 d5 5.Nxd5 Bg4+ 6.Nf3 Nc6 7.Nxc7+ Kd8 8.Nxa8 Ne5 9.h3 Bxf3+ 10.gxf3 Qg3 11.d3 Qxf3+ 12.Ke1 Qg3+ 13.Ke2 Qf3+ 14.Ke1 Qg3+ ½–½

Black has 3...Nc6 of course.

Please check Harding's column on chesscafe, I believe two or three years ago  had an analysis of this gambit.

the masons gambit is not only playable, but it offers white a great + out of the opening if you understand the lines and the concepts.  i wouldnt worry about being tortured by black in this line.  it gives more than equality against Qh4+

Thank you very much for answearing me. In the games I went trough came also Qh4, then Ke2 then d5 or d6 in both cases with unclear fights. Is there a book which coveres this?

I replayed some old games including the maision gambit (1e4 e5 2f4 ef 3Nc3)

I am sure this leads not to a white advantage. But I am wondering if this old opening is playable (unclear position) or if there is a refuation (black is better).

I felt a bit in love with the idea let your openend attack, let torture me in the middle game and ending up after a while of suffering  in a better ending.