I have read the paper (citing three tanks) a few times and while interesting, it could be improved upon. First….there are too many variables….fish species, water type, tank size, substrate makeup, food type and food amounts, filtration, plants vs no plants, temperature, and (unless I missed it) no mention of pH and KH which play in integral part of the nitrogenic process and the sampling size was too small. Learning about bottled bacteria and all is important, but test it all out maintaining other factors as constants….same fish, bioload factors, temperature, same pH and KH and GH, same food/food amounts, substrate, filtration, and tank size. Then one can compare the nitrogen cycle factors/types successfully because all the other factors are constant as opposed to one tank having a higher temp, lower pH, heavier bioload etc than another tank.
I am (respectfully) merely trying to get this sorted out with results that can be applied generally. As I see it, one could set out to do what was done with tank 1 but doing so with a much higher pH, higher bioload, for example, could greatly alter the results. All three of those tanks were not that much alike (had too many aforementioned variables) imho and thus any conclusions based on them is weak at best….especially regarding bottled bacteria.