To examine the size—dependence of prey vulnerability, replicate groups of hatchling Nucella emarginata (= Thais emarginata) were held in cages in the laboratory with five size categories of barnacles (Balanus glandula and Chthamalus dalli). By measuring the diameters of both successful drill holes and unsuccessful drill attempts, and by using the close correspondence observed between shell length and drill hole diameter, a detailed picture emerged of the ability of hatchlings (1.1—5.3 mm shell length) to attack and consume barnacles of various sizes (1.0—6.0 mm opercular diameter). These data yielded some important general insight into the shape of the vulnerability function and the nature of the size race between predator and prey. They also revealed several interesting features about the size—dependence of this particular predator—prey interaction. For three size classes of hatchling (<3.5 mm), the vulnerability of barnacles (percent successful attacks) decreased roughly sigmoidally with increasing barnacle size. The sigmoidal shape of this vulnerability function seems likely to be a general feature of size—dependence of prey vulnerability, because the size at which prey achieve an escape in size is unlikely to be discrete. Furthermore the recognition that this vulnerability function is sigmoidal in shape suggests a theoretically sounder descriptor for the maximum size of prey vulnerability ("critical size" of Vermeij): the median vulnerable size (SV50). Another intriguing feature of these data was the relationship between barnacle SV50 and hatching size. Regardless of whether length or body mass were compared, barnacle SV50 increased isometrically with hatchling size. In other words, for the size—ranges examined in this predator—prey system, neither prey nor predator appeared to achieve a disproportionate advantage from an increase in size. The relation between prey SV50 and predator size, on log—transformed axes, provides a convenient way of summarizing the size race. This study also revealed a number of interesting features about the feeding biology of hatchling N. emarginata. Surprisingly, even the smallest hatchlings (1.4 @+ 0.0222 mm shell length) were able to consume at least a few of the largest barnacles offered (6.0 mm opercular diameter), hence no discrete upper size limit of prey was observed. Attacks on barnacles of this size, however, were successful <10% of the time. By a size of 5 mm shell length, hatchlings were able to consume nearly all sizes of barnacles with at least 50% success. These data also revealed rather nicely that the apparent preference of smaller barnacles by hatchlings when given a choice, was actually an artifact of differential attack success: although attacked with equal frequency, larger barnacles were less likely to be drilled successfully. The average growth rate of hatchlings declined significantly with increasing barnacle size due in large part to a decrease in attack success, but the variation within cages was large. The attack behavior of hatchlings depended upon both hatchling size and barnacle size. The frequency of attack at sutures between skeletal plates vs through skeletal plates, and the frequency of attack in the opercular vs the lateral region of barnacles, both increased with increasing hatching size and decreased with increasing barnacle size. The increased frequencies of sutural and opercular attacks appeared to be adaptive because attacks by hatchlings at these locations were significantly more successful; however, other changes with size were not associated with differences in attack success.