Seeing shapes in clouds: the fallacy of deriving ecological hypotheses from statistical distributions