Rosy:
My apologies, the structure of this post may be a little confused as I had to go to a class part way through writing it and haven't time to start from scratch and make it into comprehensible prose(!!)
I think I've made all the points I intended to, however!
If the primary system for moving water up trees is this convection type system you're proposing, how do the sugars get *up* the trees to the ends of the branches for leaf formation in the spring? According to your model, if there aren't any leaves yet how does the flow get started and worse how does it draw more sugars (and amino acids and whatever else it needs) up than it drops down (which it must in order to construct new leaves)? It's got to use active transport in the phloem.
OK, my understanding of the current model (and I'm in no sense a plant scientist).
In the leaf cells, sugar is produced and water is lost by evaporation from the leaves.
Sugars are transferred by (mainly passive) transport (depending on the concentrations) into the phloem. Given the sugars are already (since they're made in the leaves and moved to other parts of the plant) moving down a concentration gradient, there is no reason for more water to follow them across the cell.
Loss of water from the leaves results in water being drawn up from the roots via the xylae(osmosis).
The sugars want to move to a position of lower energy/higher entropy (and so to places where there is less sugar already). There *is* an energy gain in going downwards, yes, but as Dave points out it isn't actually very big if you're losing a whole load of water at the top. Your Brixham experiment depends on using the weight of the water coming over the top of the loop to draw the water below it up. In order to produce any energy at all the salt/sugar solution has actually to move downwards, which in your model it can't do unless the water which it pulls up follows it straight back down the opposite tube. Indeed, as was pointed out by EL Hemetis, there is before us the evidence of plants quite happily growing with their leaves below their roots. I'm far more convinced by the idea that that concentration effects dominate.
Where is the need for a "simpler" explanation? The current model doesn't strike me as any more complicated than yours (and you have yourself shown that provided there's a sufficient upwards "pull" water can sustain the "tension" required), given that the mechanisms I've outlines are undoubtedly present in generalised cells and therefore presumably in trees.
Amber really is just the fossilized stuff. The "amber" people harvest is probably copal, which I *think* is a form of resin.
www.emporia.edu/earthsci/amber/copal.htm
Also, the falling through the floor thing... if I eat too many doughnuts I may fall through the floor. But only by breaking the floor. The only way of applying the same argument to sugar solution is to say that the sieve plates between phloem cells rupture. Which is OK unless the tree's ever going to want to move nutrient solutions "uphill", which is going to require (energy expensive) active transport because as you so rightly say, all other things being equal heavy solutions (and indeed rocks) want to move downwards with gravity *if the thing they're resting on can move out of the way to allow this*.