The problem can easily be understood. Compare it to opening a pic of, say, 100x80 pixels and blowing it up to full screen. You will see pixels. The quality will deteriorate. That same thing can happen with UV. When you model something, as you well know, you get polygons whose size and placements are decided by the points on their corners (vertices). An uv map of your object has the same number of polyons and points/vertices as your model, and each point corresponds with one on the polygon mesh. U corresponds with the x-value, and V with the Y value. (z, the depth, is in this case not used). Now when uv coordinates are flattened out on a map, some distortion occurs. Like flattening the earth and seeing the continents on a map: some continents are larger than they really are, and some are smaller. In your case, some polygons on the map can be much bigger than they are, and some smaller. When now, on your map, you apply your bitmap image, the larger uv polygons will get a larger piece, and the smaller ones a smaller piece of your bitmap. When seen on your object, those polygons have the same size, yet some will have more of the bitmap texture than others. This is, simplified, the case. The remedy is more complex. If you own Bodypaint, you can larn to work with it. If you don't, and the invstment is too large, you may want to get UVMapper, which is a very good app. In there, you can move points of your uvmap, you can relax certain portions so that the distortion will become less.

I have tried prudent planning long enough. From now I'll be mad. (Rumi)