Spider orb webs are dynamic, energy absorbing nets whose ability to intercept prey is dependent on both the mechnical properties of web design and the material properties of web silks. Variation in web designs reflects variation in spider web spinning behaviours and variation in web silks reflects variation in spider metabolic processes. Therefore, natural selection may affect web function (or prey capture) through two independent and alternative pathways. In this paper, I examine the ways in which architectural and material properties, singly and in concert, influence the ability of webs to absorb insect impact energy. These findings are evaluated in the context of the evolution of diverse aerial webs. Orb webs range along a continuum from high to low energy absorbing. No single feature of web architecture characterizes the amount of energy webs can absorb, but suites of characters indicate web function. In general, webs that intercept heavy and fast flying prey (high energy absorbing webs) are large, built by large spiders, suspended under high tension and characterized by a ratio of radii to spiral turns per web greater than one. In contrast, webs that intercept light and slow flying prey (low energy absorbing webs) are suspended under low tension, are small and are characterized by radial to spiral turn ratios that are less than one. The data suggest that for spiders building high energy absorbing webs, the orb architecture contributes much to web energy absorption. In contrast, for spiders that build low energy absorbing webs, orb architecture contributes little to enhance web energy absorption. Small or slow flying insects can be intercepted by web silks regardless of web design. Although there exists variation in the material properties of silk collected from high and low energy absorbing webs, only the diameter of web fibres varies predictably with silk energy absorption. Web fibre diameter and hence the amount of energy absorbed by web silks is an isometric function of spider size. The significance of these results lies in the apparent absence of selective advantage of orb architecture to low energy absorbing webs and the evolutionary trend to small spiders that build them. Where high energy absorption is not an exacting feature of web design, web architecture should not be tightly constrained to the orb. Assuming the primitive araneoid web design is the orb web, I propose that the evolution of alternative web building behaviours is a consequence of the general, phyletic trend to small size among araneoids. Araneoids that build webs of other than orb designs are able to use new habitats and resources not available to their ancestors.