These make a direct appeal to the practical man, but of the scientific methods employed the field experiments are the least under control. The plants are grown under the most natural conditions of cultivation it is possible to obtain, and for that reason much value has been attached to such tests. Certainly, so far as the final practical application is concerned, open field experiments are the only ones which give information of the kind required. But from the scientific point of view one very great drawback exists in the lack of control that the investigator has over the conditions of experiment. The seeds, application of poison, &c., can all be regulated to a nicety, but the constitution of the soil itself and the soil conditions of moisture, temperature and aeration introduce factors which are highly variable. No one can have any idea of the composition of the soil even in a single field, as it may vary, sometimes very considerably, at every step. Further, no one knows the complicated action that may or may not occur in the soil on the addition of extraneous substances such as manures or poisons. Altogether, one is working quite in the dark as to knowledge of what is going on round the plant roots. It is impossible to attribute the results obtained to the direct action of the poison applied. While the influence may be direct, it may also happen that certain chemical and physical interactions of soil and poison occur, and that the action on the plant is secondary and not primary, so that a deleterious or beneficial result is not necessarily due to the action of the toxic or stimulating substance directly on the plant, but it may be an indirect effect induced possibly by an increase or decrease in the available plant food, or to some other physiological factor. Consequently great care is needed in interpreting the results of field experiments without the due consideration of those obtained by other methods. II. DETAILS OF METHODS. Many details of the sand and soil culture methods have been published by various investigators, e.g. Hiltner gives accounts of sand cultures, while the various publications issued from Rothamsted deal largely with the soil experiments. As this is the case, and as all crucial experiments have always been and must always be done in water cultures, it is only necessary to give here full details of these.