This question is aimed to TRUE Elec Engineers, not fuse or wire directionality believers.

Regarding the woofer’s excursion; That’s what happened, with 1.5 VDC. Infer what you like. I suppose, were I to dig through my shoppe’s old inventory, found a 4 or 5uF cap, and an 8 Ohm resistor, then bought another O-Scope, I could provide more details, regarding exactly what’s occurring, regarding the DC voltage spike, through the cap(s). The facts are: the tweeter blew and there’s a distinct pop, from the woofer, when connected to a C cell, through a non-polarized capacitor. That indicates there’s a transient, being passed to the driver. As I mentioned at the outset, these are experiments that can easily be done, by anyone really interested. Anyone that does conduct such: I’d like to hear their results/conclusions(particularly, if you own/use an O-Scope).

"OK- you didn’t damage the tweeter with a 0.1uf cap, but you did with a 33uf cap. The 33 uf cap allows for much lower bandwidth of information through the tweeter, which in most cases only handle a few watts. My surmise is the 33uf cap made for too much excursion." There is no, "much lower bandwidth of information", that could possibly proceed from a battery. Also- that tweeter blew when the multimeter was taken out of series. I never tried it, using the .01uF, without the multimeter. I may just dig a tweeter out of my old inventory, to(perhaps) sacrifice. I had a box of EV Interface A & B, paper coned tweeters, for doing EV’s warranty work(REAL cheap), back in the early Eighties).  That'll have to wait until the temps in my garage, get above 0 Degrees(F).  The premise that I’ve been testing, is that NO DC voltage would get past a non-polarized capacitor. At least, that’s what’s been proffered in this thread.

Back to back MOSFETs suggested here will work but they need a 10V floating voltage source or some photovoltaic coupler to a DC detection circuit. These have time delays. The tweeter should be able to handle a short single spike. Average power will be low.
Power requirements for a tweeter is very low so driving from a 100V source kinda don't make sense. Bi amping using low power amp for tweeters does.
Tube amps have transformers isolating from DC and typically sound good.
A normal mechanical relay has a delay time as well, like 8 - 20 ms  so some energy will reach the tweeter. I would think there is a cap in series with the tweeter, that will block DC. A small relay could be placed after the cap maybe even a reed relay.
As some have posted here, a fast crowbar circuit is used to blow a fuse where extreme protection is needed.
If a power amp has glitches that puts DC on the output it should be fixed or else replaced. 
Additionally, HF energy can pass the speaker high-pass filter and some protection for the tweeter is needed here.  I have burned up tweeters with amp stability problems. Perhaps limit the max frequency the tweeter is exposed to. 
DC on the output of audio amps is a problem that was solved 40 years ago, as I can recall. This really should not be an issue.