This method has the advantage over water cultures in that the environment of the plant roots is somewhat more natural, but on the other hand the work is cumbersome and costly, while the conditions of nutrition, watering, &c., are less under control than in the water cultures. Sand cultures represent an attempt to combine the advantages of both soil and water cultures, without their respective disadvantages. Generally speaking perfectly clean sand is used varying in coarseness in different tests, and this is impregnated with nutritive solutions suitable for plant growth. The sand is practically insoluble and sets up no chemical interaction with the nutritive compounds, while it provides a medium for the growth of the plant roots which approximates somewhat to a natural soil. It is probable, however, that a certain amount of adsorption or withdrawal from solution occurs, whereby a certain proportion of the food salts are affiliated, so to speak, to the sand particles and are so held that they are removed from the nutritive solution in the interspaces and are not available for plant food, the nutritive solution being thus weakened. The same remark applies to the poisons that are added, so that the concentration of the toxic substance used in the experiment does not necessarily indicate the concentration in which it is presented to the plant roots. On the other hand, undue concentration of the solution is apt to occur on account of the excessive evaporation from the surface of the sand. The sand particles are relatively so coarse in comparison with soil particles that the water is held loosely and so is easily lost by evaporation, thus concentrating the solution at the surface, a condition that does not apply in soil work. With care this disadvantage is easily overcome as it is possible to weigh the pots regularly and to make up the evaporation loss by the addition of water. 3. _Soil cultures in pots._ In this case the conditions of life are still more natural, as the plant roots find themselves in their normal medium of soil. But the investigator has now far less control, and bacterial and other actions come into play, while the nutrients and poisons supplied may set up interactions with the soil which it is impossible to fathom. This method is useful in the laboratory as it is more convenient for handling and gives more exact quantitative results than plot experiments. Also the pots can be protected from many of the untoward experiences that are likely to befall the crops in the open field. The conditions are somewhat more artificial, as the root systems are confined and the drainage is not natural, but on the whole the results of pot experiments are very closely allied to those obtained in the field by similar tests. 4. _Field experiments._