Reciprocal transfers may be made by means of seed or plant. Since the experiment is a complex one, all the care possible should be taken to make sure that the plants become established in the reciprocal situations, and consequently, it is often advisable to transfer both seeds and plants. Reciprocal transfer is of paramount value in solving the problem which bog plants present. A slight modification of the method makes it possible to obtain experimental evidence of the polyphyletic origin of species in consequence of adaptation. In an experiment mentioned elsewhere, the transfer of _Kuhnistera purpurea_ to the area occupied by _K. candida_, and vice versa, is designed to show whether one has been derived from the other. If the two species are moved into an area which contains more water than that usually occupied by _K. purpurea_, and less water than is found where _K. candida_ habitually grows, the resulting modifications will throw much light upon the origin of polyphyletic species. In this connection, it hardly needs to be pointed out that this simple transfer of a species to several separated areas of a new habitat may often furnish complete proof that a new form may arise at different times, and at different places. [Illustration: Fig. 47. Control ecad of _Ranunculus sceleratus_, holard 10% (50 cc.).] =194. Modification of the habitat.= Efficient changes in the habitat are brought about by increasing or decreasing the water-content, or by varying the light intensity between sunshine and the diffuse light of deep forests. Humidity can not well be regulated except in so far as it is connected with water-content. Since its effects merge with those of the latter, its modification is unnecessary. An increase in water-content is readily brought about by irrigation. A stream may be dammed and its water allowed to spread over the area to be studied, or the water may be carried to the proper place by deflecting the stream or by digging a canal. The construction of earth reservoirs makes it possible to obtain almost any per cent of soil water by varying the size of the reservoir or the height of the wall or bank. Near a base station, such as Minnehaha, where there is a simple system of water-works, the experimental area may be watered whenever desirable by means of a hose. Water-content may be readily decreased by drainage, or by the deflection of a stream. When such means are not available, as in the case of extensive marshes, hummocks may be used or constructed, and the soil blocks containing plants placed upon them. By the use of sand or gravel, the water-content of mesophytic areas can be reduced in a similar manner, or by surrounding the plant _in situ_ with either of these soils which hold little water. In meadows, especially, the addition of a large quantity of alkaline salts decreases the amount of available water, while the holard may be reduced by denuding the soil about the plants concerned.