Three strains of the bar-eyed mutant of Drosophila melanogaster Meig have been reared at constant temperatures over a range of 15-31 degrees C. The mean facet number in the bar-eyed mutant varies inversely with the temperature at which the larvae develop. The temperature coefficient (Q(10)) is of the same order as that for chemical reactions. The facet-temperature relations may be plotted as an exponential curve for temperatures from 15-31 degrees . The rate of development of the immature stages gives a straight line temperature curve between 15 and 29 degrees . Beyond 29 degrees the rate decreases again with a further rise in temperature. The facet curve may be readily superimposed on the development curve between 15 and 27 degrees . The straight line feature of the development curve is probably due to the flattening out of an exponential curve by secondary factors. Since both the straight line and the exponential curve appear simultaneously in the same living material, it is impractical to locate the secondary factors in enzyme destruction, differences in viscosity, or in the physical state of colloids. Differential temperature coefficients for the various separate processes involved in development furnish the best basis for an explanation of the straight line feature of the curve representing the effect of temperature on the rate of physiological processes. Facet number in the full-eyed wild stock is not affected by temperature to a marked degree. The mean facet number for fifteen full-eyed females raised at 27 degrees is 859.06. The mean facet number for the Low Selected Bar females at 27 degrees is 55.13; for the Ultra-bar females at 27 degrees it is 21.27. A consistent sexual difference appears in all the bar stocks, the females having fewer facets. This relation may be expressed by the sex coefficient, the average value of which is 0.791. The average observed difference in mean facet number for a difference of 1 degrees C. in the environment in which the flies developed is 3.09 for the Ultra-bar stock and 14.01 for the Low Selected stock. The average proportional differences in the mean for a difference of 1 degrees C. are 9.22 per cent for Ultra-bar, and 14.51 for Low Selected. The differences in the number of facets per degrees C. are greatest at the low and least at the high temperatures. The difference in the number of facets per degrees C. varies with the mean. The proportional differences in the mean per degrees C. are greatest at the lower (15-17.5 degrees ) and higher (29-31 degrees ) temperatures and least at the intermediate temperatures. Temperature is a factor in determining facet number only during a relatively short period in larval development. This effective period, at 27 degrees , comes between the end of the 3rd and the end of the 4th day. At 15 degrees , this period is initiated at the end of 8 days following a 1st day at 27 degrees . At 27 degrees this period is approximately 18 hours long. At 15 degrees it is approximately 72 hours long. The number of facets and the length of the immature stage (egg-larval-pupal) appear related when the whole of development is passed at one temperature. That the number of facets is not dependent upon the length of the immature stage is shown by experiments in which only a part of development was passed at one temperature and the remainder at another. Temperature affects the reaction determining the number of facets in approximately the same way that it affects the other developmental reactions, hence the apparent correlation between facet number and the length of the immature stage. Variability as expressed by the coefficient of variability has a tendency to increase with temperature. Standard deviation, on the other hand, appears to decrease with rise in temperature. Neither inheritance nor induction effects are exhibited by this material. This study shows that environment may markedly affect the somatic expression of one Mendelian factor (bar eye), while it has no visible influence on another (white eye).