Alt Bike Culture - Lever drive bike idea.

I had just thought up some ideas for replacing a chain with levers - the gears would be replaced by cranks.
http://i5.photobucket.com/albums/y158/XYZZY2000/leverbike.jpg
Ideally, you'd just want a single rod to go from the pedals straight to the crank that would rive the rear wheel, since this would be the most efficient and light, but it would only be possible to obtain a 1:1 drive ratio with this method unless there was some sort of internal gearing in the rear hub.
The other idea, on the left uses a rocker arm, allowing the forward connecting rod to travel a greater distance than the rear one - thus giving an increase in "gearing". The farther apart the connecting rods are mounted on the rocker arm the higher the gearing would be, the closer together the lower they would be.
Unfortunately, for this to be variable in any sense of the word, the rocking arm would have to be curved so as to make up for the fact that the connecting rods can't stretch as you move them apart from eachother - a suppose that'd be possible - but I haven't a clue as to how one would actually be able to "shift" the rods while on the bike, anyway fabicating a curved rocker would probably be well beyond my skills.

Anybody have any ideas on what might make such a silly idea workable?

There's a similar invention here:

http://www.sciencenewsforkids.org/articles/20030827/Feature1.asp

And yes, your idea looks workable.

The other idea, on the left uses a rocker arm, allowing the forward connecting rod to travel a greater distance than the rear one - thus giving an increase in "gearing". The farther apart the connecting rods are mounted on the rocker arm the higher the gearing would be, the closer together the lower they would be.

Maybe I'm looking at it wrong, but wouldn't that design also be a one to one ratio? Either that or it would just bind up.

No, you see because of the rocker acting between the two connecting rods, the forward rod can be mounted to a longer crank than the rear rod. So the front rod could travel say twice the distance (or half as fast) as the rear rod due to being mounted on a crack that is twice as long as the crank for the rear rod. The rocking arm allows them to be mounted indirectly eliminating the binding that would occur if two cranks of different sizes were connected directly by a single rod. The rocker is connected at one end to the frame, so the farther away from that point a connecting rod is mounted, the greater the distance it can travel, the closer it is mounted to the point the less distance it can travel, this way a small crank can be fitted the rear wheel, and a larger one to the pedal crank - in effect doing the same job as a small cog on the rear wheel and a large chainring on the pedal crank.

Look up 4 bar linkage. If you can adjust the linkage length in real time you would have a continuously gearing system. Pedal Cars would be a starting point. See Champiot rowed quadracycle.

You'd need a totally different rear drops and axle set up. Instead of the normal axle clamped in the drops that doesn't rotate, you'd need a rotating axle (or spindle) similar to the one that connects the cranks.

Also, no matter what length the various interconnecting rods are, it looks like one turn of the crank yields one turn of the wheel, so the only way I can see to adjust gearing is via hub gears.

Take a look at locomotive wheels....

Well the levers don't really change gearing, it's more like they're making the pedal cranks longer without actually putting on longer pedal cranks.
And dropouts, they could be exactly the same, the crank that drives the wheel would run through a sleeve that'd be fixed into the right hand drop out. It'd probably require a roller bearing too.
OR... I could go back to the original single con rod idea, only have it geared to the the rear hub and geared to the pedal cranks, then a conventional rear hub could be used and the gearing could be solved, but this probably kills anything resembling efficiency that the bicycle may have otherwise possessed.

unless you're using a single sided swingarm setup like in some motorcycles your arm won't make it around the axle. you'd have to over engineer the single sided wheel mount which would make the whole design much more complex than necessary, not to mention the weight issues. In engineering there is a rule called the K.I.S.S. rule. It stands for Keep it Simple, Stupid. meaning that the simplest solution is most often the best one. Less to go wrong, Less to break, and easier to fix. I'm not trying to shoot down your interesting Idea, but It could really go no further than a novelty thing.

unless you're using a single sided swingarm setup like in some motorcycles your arm won't make it around the axle. you'd have to over engineer the single sided wheel mount which would make the whole design much more complex than necessary, not to mention the weight issues. In engineering there is a rule called the K.I.S.S. rule. It stands for Keep it Simple, Stupid. meaning that the simplest solution is most often the best one. Less to go wrong, Less to break, and easier to fix. I'm not trying to shoot down your interesting Idea, but It could really go no further than a novelty thing.
You could make a cardboard model, just to verify that the system would work at all.

Regarding KISS, a typical bike has about 120 moving parts in the drivetrain. About 115 of these parts need to be replaced every 2000 miles depending on type of bike and usage. So if someone could figure out a way to reduce the number of parts to only 3 or 4, it would greatly decrease complexity. Weight wouldn't be that much more substantial, and with technological refinements would undoubtedly come down.

Of course, the fact that it hasn't been done in the last 120 years suggests that it probably can't be ... reasonably.

I was not really looking at this as a serious idea, so much as just producing an interesting and novel bicycle as a conversation piece. I think that it would probably give pedalling a very unusual sensation, since instead of constantly pulling a chain around, you'd be swinging weight back and forth, unless you ballanced every rod perfectly. I have absolutely no doubt that it'd work. But I seriously doubt it'd be practical in even the least to most people.