Ultimately, the HMBC spectrum relies on the coupling constant between the carbon and proton atoms. If you don't see a correlation you would expect, then the coupling constant is not in the suitable range. There is a lot of overlap in the coupling constants of two- and three-bond C-H couplings, and quite often you see both, but not always.

As for your second question, The coupling is still there, even in symmetric molecules, and what you see are the couplings from C3 to H5 and from C5 to H3.

Finally, the correlation from C8 to H8c is a one-bond coupling that is split by the C-H coupling constant. These occasionally show up in HMBC spectra, and there are pulse programs that use low-pass J-filters to remove such correlations (for example, hmbcetgpl3nd on bruker systems, used by default by the parameter set HMBCETGPL3ND)

Jani Rahkila Much appreciated.

Something which I am also observing is that in para substituted systems, such as 4-aminophenol as an example, the carbon bonded to the amino group couples with the meta hydrogens instead of the ortho hydrogens and the same goes for the carbon bonded to the phenol group. My guess is that since the two bond correlations are very strong (especially because there are two equivalent hydrogens to which the carbon can long range couple) the peak gets filtered out along with the single bond correlations.

Dear Roderick,

an addition to Jani Rahkila 's instructive answer:

In benzene derivatives, 2J(C,H) is usually much smaller (2 Hz) than 3J(C,H) (7 Hz). That's the answer to your question "Why is it that in the attached image: H4 is not coupling with C5, C6 if they are separated by two bonds?". H-C HMBC experiments are usually optimized to detect couplings of 8 to 10 Hz, so you will find the correlations 4-H/C6 or 2-H/C6 with a much larger intensity than 4-H/C5. If you select for other coupling constants, the picture will change.

And for your other observation (para-aminophenol): You are sure that the intensive correlation signals repesent the 2J connectivities? I really doubt about that.

Appreciated and thanks. It makes sense. Wolfgang Baumann do you have an idea if four bond correlations are observed? Also, is the coupling between a quaternary carbon and the hydrogen attached to the vicinal carbon larger than between a non-quaternary carbon and the hydrogen attached to a vicinal carbon (in an aromatic system)?

Dear Roderick,

for illustration I have attached a copy of a spectrum (an undisclosed benzene derivative). According to my experience, it does not matter whether a correlation to a "protonated" or an "unprotonated" carbon is observed, coupling constants are similar, but I do not have exact numbers at hand. Such data are rarely evaluated and even less frequently published or documented. From the picture you can see the correlation signals belonging to the "arrowed" hydrogen position. Both 3J (the yellow boxes) give strong response: E is a CH group and H is a quart. C. 2J (the red box) gives weaker response: D is a CH (not even seen) and F is quart. In some cases I also see correlations through four bonds: here the one in the green box.

Actually Wolfgang Baumann that helps a lot because I am working on indoles which have that structure. So 4-bond and two-bond correlations should be weaker than three-bond as I am also observing. Perfect.

Thanks again