Closer examination, however, shows that this division into three classes is too rigid. Now that experiments are more refined it has become evident that no such simple grouping is possible. It has been found that typical nutrient salts are toxic when they are applied singly to the plant in certain concentrations, the toxic power decreasing and the nutritive function coming into play more fully on the addition of other nutrient salts. For instance, Burlingham found that the typical nutrient magnesium sulphate in concentrations above _m_/8192 (_m_ = molecular weight) is toxic to most seedlings, the degree of toxicity varying with the type of seedling and the conditions under which growth takes place. It will be shown in the following pages that even such a typical poison as boric acid may, under suitable conditions, increase plant growth just as if it were a nutrient. A review of the whole subject leads one to conclude that in general both favourable and unfavourable conditions of nutrition are present side by side, and only when a balance is struck in favour of the good conditions can satisfactory growth take place. As indicated above, experiments have shown that the very substances that are essential for plant food may be, in reality, poisonous in their action, exercising a decidedly depressing or toxic influence on the plant when they are presented singly to the roots. This toxic action of food salts is decreased when they are mixed together, so that the addition of one toxic food solution to another produces a mixture which is less toxic than either of its constituents. Consequently a balanced solution can be made in which the toxic effects of the various foods for a particular plant are reduced to a minimum, enabling optimum growth to take place. Such a mixture of plant foods occurs in the soil, the composition of course varying with the soil. While the earliest observations set forth the poisonous action of various substances upon plants, it was not long before investigators found that under certain conditions these very substances seemed to exert a beneficial rather than an injurious action. The poisons were therefore said to act as “stimulants” when they were presented to the plant in sufficiently great dilution. This stimulation was noticed with various plants and with several poisons, and a hypothesis was brought forward that attempted to reconcile the new facts with the old conceptions. Any poison, it was suggested, might act as a stimulant, if given in sufficiently small doses. It will be seen in the following pages that this is not universally true, such substances as copper, zinc, and arsenic failing to stimulate certain plants even in the most minute quantities so far tested.