We first counted the number of species in each family among
the species used to treat gastrointestinal diseases. A high number of
species in a given family may, however, have two explanations: the
given plant family might have properties making it a valuable
source of medicine, or it might only be common in the area. To
explore these two possibilities, we tested if some botanical families
had more species used in gastrointestinal diseases than expected
from the total number of species in each family in the flora of Nepal
(Press et al., 2000). Specifically, we tested whether the plants used
in gastrointestinal diseases were random subsets of all vascular
plants and all medicinal plant species regarding identity to botanical
family, following the method described in Rokaya et al. (2012b).
For this, we performed permutation tests by drawing random
samples of plants from the whole flora of Nepal and assigning
them to plant families. We asked whether the number of species
used in gastrointestinal diseases in a given family could result from
a random sampling from the whole flora or from the medicinal flora
of Nepal. For each family, we derived an empirical P-value, which
indicated the probability that the number of species of a given
family used to treat gastrointestinal diseases could be due to
random sampling from the total flora or medicinal flora of Nepal.
We performed 10,000 permutations using