The author's own divice, as mentioned in part Ⅰ, a method of substitution of the dye potassium mercuric iodide complex formed in the supravitally stained cells with mercuric sulfide complex, gave a good result to retain the dye granules in situe, revealing the selective staining of mitochondria with Janus green B. But the granules appeared too opaque to see the detailed inner structure of mitochondria. Then, the author tried to retaine the potassium mercuric iodide dye complex through the dehydration process by using a specific dehydration media for electron microscopy. In vitro experiment on the precipitants, the mercuric potassium iodide dye complexes, proved that this complex is dissolved easily in pure ethanol, but is not dissolved in 50% ethanol. On the author hand, the complex is not dissolved in ethyl-ether, but dissolved ether-ethanol mixture (1.1). The test done by lowering the ethanol contents in ether proved that the dye does not dissolved in the ether containing 20% ethanol. On the basis of these observations the author tried to retain the complex through dehydration and succeeded in observing the mercuric-dye complex in the cells under electron microscope by making the sections from the cells dehydrated through 50% alcohol, ether containing 10% of pure ethanol and then pure ether, imbibed with ether-methacrylatemonomer and methacrylatemonomer, and embedded in mathacrylate by the aid of benzoyl peroxide. The electronmicroscope observation on the cells stained supravitally with Janus green B, fixed by the method deviced by SENO, dehydrated and embedded by the method just mentioned, proved that the minute structure of the cells are kept as well as in those fixed with 1% osmic acid and dehydrated by the routine method. In the series of observation on the cells treated with Janus green B in the different period of staining, the processes of the invasion of dye into mitochondria and the deformation of the inner structure have bean traced. After a short period of staining, the dye has acted as scarecely to increase the opacity of mitochondria but a marked change in they structure occured by the ring formation at one end of mitochondria. More deeply stained ones the appear very dark but in the dark area some opaque droplets are recognized. In this case, too, the less opaque pole were seen at the end of mitochondria, in which the structure of cristae were observed. These observations show that by the action of dye element of mitochondria seems to be segregated into 3 parts, less affine, modercetely affine and strong affine elements to me dye, with the destruction of the inner structure.