Let me also give some insight on this topic. I have experience with sending tobacco samples to a lab for nicotine tests. That is, nicotine testing is not the main objective, but it does get tested. I can only give an outlay on the results of these tests, not explanations as to how these results come about. We buy tobacco from a large number of farmers and we categorize them into areas. We then make samples per area for tests. So each sample contains a few leafs from all farmers in that area. Per area there are two tests, one for middle plant positions, the other for top plant positions. Remember that all tobacco is only a single variety, although there may be various sub-varieties. I have here 20 tests from 10 areas. 10 middle positions tests and 10 top position tests. Now assuming that we should look at pairs of tests (per area), on 5 occasions the higher plant position tests for a lower nicotine level than the middle plant positions. On 3 occasions the top plant position tests higher than middle plant position. One of these I would disregard as the population size is extremely large and the samples of middle and top plant position hardly relate. On 2 occasions top and middle test exactly the same. Now I would say that is pretty random. Nicotine levels vary between 0,75% and 2,10%. Now maybe more interesting is that I have before me one more sample, which comes from tobacco that was grown organic. Only a bit of fertiliser but no pesticides. That sample tests 0,46%. TO me that was surprising as I had assumed that a plant with no pesticides would generally make more nicotine to protect itself from the outside world. I was wrong, apparently.