It's an interesting article tough a tad flawed at some points.

The Caapi and P.Vridis are mixed up, Caapi is the MAOI and P.Viridis is the DMT containing plant.

Also the assumpsion of the LD50 for humans is of course a guess to be on the safe side. We never know for sure if an LD50 for rodents also applies to humans so we lower the assumed LD50 x fold. I'm not sure when this research was done but a bit more recent research would have shown that the LD50 for harmine in the article is way too low to be true. If we follow the divide by 20 law of the article the LD50 for harmine would be 0.1mg/kg! A normal dose of harmine in huasca is about 200mg and we aren't dropping dead one half of the time, even without the 20x divider it's an unreasonable low LD50. Also, a DMT-dose higher then 112mg/70gk is pretty normal and we don't drop dead the other half of the time.

After reading that sentence about 10 times I finally got it. The 520 is a typo and should be 560 just like the second fivehundred number that gets divided by 27 (8x70=560). Also they don't mean about 20mg but 20 times! That sounds better, since 20mg difference between a normal and a lethal dose is way too risky.