Excuse my ignorance, but, I have always thought (probably wrongly) that a 1.6x crop factor on a camera gave your lens more 'reach' than a 1.3x or a full frame camera.
I must have mis-interpreted what I have read on other websites.
When you stick a 400mm lens on the front of a camera the size of the subject projected onto the sensor is going to be the same, whether the camera is full frame, 1.3X crop or 1.6X crop.
e.g. Take a look at the image here -
http://digital-photography-school.com/wp-content/uploads/2008/08/full-frame-crop-factor.jpg. It's not my image so I won't attach it. Please follow the link. Now, imagine that in the centre of that photo you have your subject - a bird, rabbit, person, whatever. The size of the sensor will make not a scrap of difference to the size of the subject as recorded at the sensor. If the subject occupied 10mm of height on a 40D sensor it would occupy 10mm of height on a 1D3 sensor and 10mm of height on the sensor of a 5D2 as well. There is no extra magnification achieved, or extra reach, through having a smaller sensor. You simply end up recording less of the scene as a whole, which if you are short on focal length is absolutely fine. You save money on the body and end up with smaller files that aren't full of pointless pixels that you will need to crop out later.
So, as far as sensor size is concerned, rather than thinking of the cropped bodies as offering more "reach", think of it as the larger sensor bodies including more of the scene. The actual subject recorded on the sensor is the same physical size (in mm, not pixels) regardless.
Where the difference in "reach" does lie is in pixel density. Quite simply, if the image of your subject projected onto the sensor is the same physical size, regardless of sensor size, then the camera with the highest pixel density will place more pixels onto your subject. If you have more pixels on your subect then you should have more fine detail recorded. Of course, there is no free lunch, and generally it is true that smaller pixels, which cannot gather as much light as larger ones, will each be a litle noisier. But add the pixels together to form the image as a whole (instead of pixel peeping) and you should see more detail in your image.
If you really want to see better IQ then you do need a larger sensor, and longer glass to fill it, so that you project a larger image onto the larger sensor and end up with far more light captured and a superior image all round. Real "reach" comes from using longer glass. For those of us on a budget we have to make up for lack of real reach by having cameras with higher pixel densities.
Even with that said, you may find that due to factors such as blur, shake, noise, misfocus, soft lens etc. that you cannot gain much advantage, if any, from higher pixel densities, because unless your capture is perfect, all they will do is more accurately record the flaws in your image capture.
To put this another way, I find that for BIF I will typically get just as much "reach" from my 1D3 as from my 50D. This is because through combinations of blur/shake/misfocus (even if only by an inch or two), noise and so on, all that the 50D really does is to reveal the softness of the capture unless my photography is perfect, which it mostly isn't. The larger pixels of the 1D3 tend to obscure slight flaws in sharpness, as well as being individually less noisey, and that gives me files that are easier to work with.
But, for perched birds, where I do not need super high shutter speeds, or high ISOs, and when I can nail focus perfectly, and when I can fully control camera shake, the 50D can pick out details that the 1D3 will miss. Thus, in simple terms, I find the 1D3 to be the better camera for things that move fast, or anything in poor light, and the 50D is better for things that barely move, and are bathed in good light. Horses for courses - right tool for the job. If you're on a budget and want something that falls between those two extremes then the 40D seems the ideal compromise. If only it had AF microadjustment.