So thanks to everyone who has posted replies so far.
As Farm Pilot has asked the obvious question I'll try to explain what I did and my thought process. I'm willing to accept that I may have made the incorrect decision if popular opinion disagrees with my actions. As I said in my original post, I'm trying to learn from this and hope that maybe this discussion will help someone else one day...
As Teefor Gage said, not putting yourself in the situation is not an option. Both ends of the airport head out into the ocean and there are houses all around. A bit of a buggar but that's what we have to deal with.
My option was to head for land, even though I knew I wouldn't quite make it and accept the downwind auto, given that the wind speed was no greater than 10kts. As we were already 90 degrees to the runway (x-wind remember) it only required a further 90 degree turn.
Just so you understand, this was a double vee-belt failure for reasons currently unknown and no warning light to indicate impending failure. I would never normally suggest that anyone should ever do an auto any way except into wind. However, in saying that what follows was the sequence of events plus my thoughts and actions as things happened:
Loud thud from engine bay area - ****, what was that?
Immediate left yaw, rotor rpm decay and engine overspeed - Holy crap
Enter autorotation
Immediate assessment of where we were
Any boats nearby? No - buggar
Into-wind auto into deep water or keep going right, towards shore? - Won't quite make it but shallow water and only a light tail wind.
Emergency radio call stating intentions
Manipulated throttle to see if any drive left at all - Nope, none, buggar again
Keep RRPM low'ish (90-95%) and extend glide as far as possible before flare
Instruction to student to open door before impact
Flare, flare, flare, holding until airspeed gone and groundspeed minimal
Deliberately let tail touch water first then levelled and raised collective simultaneously with gentle water contact
Helicopter immediately filled with water, rolled uncontrollably to the right, rotors hit and stopped, student exited and I climbed up over the now jammed collective as the helicopter sank.
Time from impact to completely submerged - approx 5-10 seconds.
Now, as far as I was concerned, having undergone Helicopter Underwater Evacuation Training and understanding how quickly a helicopter sinks and how easy it is to panic and get disoriented, as soon as the emergency was in progress my student stopped being a student and became a passenger who's life might depend upon my actions as PIC
I decided on the shallow water option because most people who die in a ditching do so as a result of drowning - not from the impact. Ok, you can be equally successful at drowning in 10' of water as you can be in 60' but I didn't want to be trying to dive down 60' to the wreckage to try and extricate my 'passenger' if for whatever reason they hadn't managed to get out.
So, any other comments anyone would like to make? Suggestions? By all means be critical but please be constructive at the same time. Thanks!