Comparative studies were carried out regarding the effect of temperature on the plasma membrane potential recorded from "controller cells" (S-potential producing cells, isolated fish retina) and from "conductor cells" (spike conducting neurons, isolated frog dorsal root ganglion). The effect of temperature changes was pronounced and immediate on the controller cells but was small on the conductor cells, the results well agreeing with those presented in the preceding paper as to the effects of anoxia, CO 2 and NI-I~. The resting potential of the horizontal cells was always hyperpolarized by heating from 20 ยฐ (room temperature) to 30 ยฐ C, and depolarized by cooling from 20 ยฐ to 10 ยฐ C, the light-induced responses being diminished at low and high temperatures. The temperature range between 18 ยฐ and 23 ยฐ C was optimal for the controller cell function, within this range the Qi0 being 2--3 (4--5 mV/C ยฐ) for the horizontal cell resting potential. The hyperpolarized resting potential of the controller cell during anoxia was much less sensitive to temperature changes. The resting potential of the conductor cells in fresh preparations was slightly depolarized by heating and hyperpolarized by cooling, this membrane behavior being opposite to that of the controller cells. A strong temperature dependence, however, was observed in the spike duration of the conductor cells. The results suggest that there exist non-synaptie cellular interactions between adjacent amacrine cells in the retina, and between the conductor cell and its satellite cells in the dorsal root ganglion. LOI~ENTE DE N6 (1947) has investigated the effect of temperature on the sciatic nerve and has given a review of earlier studies carried out on whole nerve preparations. Studies concerning the influence of temperature on the transmembrane potential of nerve fibers have been conducted by , by HODGKI~ and KATZ (1949) and by NAI~AHASHI (1962). Several workers have studied the temperature effect on the membrane potential of cells other than neurons, e.g. skeletal muscle [6,15,25], Purldnje fiber[6], Pacinian corpuscle [11,12] and electroplax [20]. The above mentioned investigations demonstrate a marked temperature dependence of the spike process and an uncertain temperature dependence of the resting potential. Opposite effects of temperature changes on the resting potential of S-potential-producing controller cells and of neurons have been demonstrated in the isolated retina by LAUFEI~ et al. (1961). Recently,