I am far from denying, however, the many important hints the natural system may afford to the philosophical student of the materia medica and to him who feels it his duty to discover new medicinal agents; but these hints can only help to confirm and serve as a commentary to facts already known, or in the case of untried plants they may give rise to hypothetical conjectures, which are, however, far from approaching even to probability.
But how can a perfect similarity of action be expected amongst groups of plants, which are only arranged in the so-called natural system, on account of often slight external similarity, when even plants that are much more nearly connected, plants of one and the same genus, are sometimes so different in their medicinal effects.
Examples of this are seen in the species of the genera impatiens, serapias, cystisus, ranunculus, calamus, hibiscus, prunus, sedum, cassia, polygonum, convallaria, linum, rhus, seseli, coriandrum, aethusa, sium, angelica, chenopodium, asclepias, solanum, lolium, allium, rhamnus, amygdalus, rubus, delphinium, sisymbrium, polygala, teucrium, vaccinium, cucumis, apium, pimpinella, anethum, seandia, valeriana, anthemis, artemisia, centaurea, juniperus, brassica.
What a difference betwixt the tasteless tinder amadou (boletus igniarius) and the bitter, drastic boletus laricis; betwixt the mushroom (agaricus deliciosus) and the agaric (agaricus muscarius); betwixt the woody stone moss (lichen saxatilis) and the powerful Iceland moss (lichen Islandicus!)
Though I readily admit that, in general, similarity of action will be much oftener met with betwixt species of one genus, than betwixt whole groups of families in the natural system, and that an inference drawn from the former will have a much greater degree of probability attaching to it, than one from the latter; yet my conviction compels me to give this warning, that, be the number of genera ever so many whose species resemble each other very much in their effects, the lesser number of very differently acting species should make us distrustful of this mode of drawing inferences, since we have not here to do with mechanical experiments, but that most important and difficult concern of mankind-health.