In attempting to use the number of chloroplasts or starch grains as a measure of response, the study should be confined to the sun and shade forms of the same species, or, in some cases, to the forms of closely related species. The margin of error is so great and the connection with light sufficiently remote that comparisons between unrelated forms or species are almost wholly without value. It has already been stated that starch is merely the surplus carbohydrate not removed by translocation; the amount of starch, even if accurately determined, can furnish no real clue to the amount of glucose manufactured. In like manner, the number of chloroplasts can furnish little more than an approximation of the amount of chlorophyll, unless size and color are taken into account. In sun and shade ecads of the same species, the general functional relations are essentially the same, and whatever differences appear may properly be ascribed to different light intensities for the two habitats. The actual counting of chloroplasts and starch grains is a simple task. Pieces of the leaves of the two or more forms to be compared are killed and imbedded in paraffin in the usual way. To save time, the staining is done _in toto_. Methyl green is used for the chloroplasts and a strong solution of iodine for the starch grains. When counts are to be made of both, the leaves are first treated with iodin and then stained with the methyl green. The thickness of the microtome sections should be less than that of the palisade cells in order that the chloroplasts may appear in profile, thus facilitating the counting. The count is made for a segment 100 μ in width across the entire leaf. Two segments in different parts of the section are counted, and the result multiplied by five to give the number for a segment 1 millimeter in width. Although sun and shade leaves regularly differ in size and thickness, no correction is necessary for these. Size and thickness stand in reciprocal relation to each other in ecads, and thickness is largely an expression of the absorption of light, and hence of its intensity. In the gravel, forest, and thicket ecads of _Galium boreale_, counts of the chloroplasts gave the following results. The gravel form (light 1) showed 3,500 plastids in the 1–mm. segment, the forest form (light .03) possessed 1,350, and the thicket form (light .002), 1,000. In these no attention was paid to the size and form of the plastids in the different leaves, since the differences were inappreciable. When this is not the case, both factors should be taken into account. Starch grains are counted in exactly the same way. Indeed, if care is taken to collect leaves of forms to be compared, at approximately the same time on sunshiny days, a count of the chloroplasts is equivalent to a count of the starch grains in the vast majority of cases. Measurements of the size of starch grains can be made with accuracy only when the leaves are