SYNOPSIS. Marine invertebrates develop in waters that extend from the poles to the equator, experiencing the full range of environmental temperature and food conditions. How selection has modified their development under the influence of these two factors has been a matter of debate. In this paper we argue that the primary influence on developmental rate is temperature, while other factors such as food availability are much less important. From existing literature we demonstrate that (1) developmental rates of both lecithotrophic and planktotrophic asteroids decrease in a similar way from the tropics to the poles, as they do also in other groups of invertebrates (echinoids, molluscs, crustaceans), and (2) rates of development at any one temperature cluster around the function describing the effect of temperature, without any relationship to egg size, suggesting that developmental rates are near the maximum for a given temperature regardless of other variables such as nutrition. We also investigated the response of development to temperature in four species of planktotrophic asteroids, one tropical, one temperate, two polar. There was limited temperature compensation among these four species, but little or no apparent ability to compensate for the retarding effects of reduced temperature within species. Arrhenius analysis of the data suggests that Q10 values for the upper region of each species' tolerance range are approximately 2, indicating that enzyme-based reactions have evolved to be closely integrated with uncatalyzed, temperature-dependent, physicalchemical processes. Values of Q10 at lower regions of the tolerance range, on the other hand, range between 9.5 and 14.7, indicative of abrupt temperature-dependent shifts in reaction equilibria, or in the organization of macromolecules and membranes. We conclude that temperature itself, rather than egg size, food, or other variables, best explains observed latitudinal differences in developmental rates in marine invertebrates.