Defence sounds are widespread and diverse amongst insects. Despite their ubiquity and variability, hypotheses explaining their functions and evolutionary origins have been understudied. My thesis focused on these topics using silk and hawkmoth Bombycoidea caterpillars as a model system. In Chapter Two I investigated why defence sounds have evolved in some caterpillars but not others by testing the hypothesis that large body size is a factor in the evolution of defence sounds. To test this hypothesis, I followed the development of defence sounds in four Bombycoidea species from hatching to pupation. I predicted that early instars would not produce defence sounds, and that within sound producing instars defence sounds would be more likely to occur in larger caterpillars. Results showed that defence sounds were absent in the first and second instar, and that they developed in the third through to the fifth instar in all species. Moreover, the onset of sound production occurred when all species were the same relative size (~1.12 g, ~26.37 mm), despite the fact that the species differed in their final instar size. I concluded that early instar caterpillars do not make defence sounds, and that there is a critical size when defence sounds develop. I further tested the hypothesis that smaller caterpillars do not have enough energy to make defence sounds, by analyzing the relationship between size and several temporal characteristics of the sounds. I predicted that smaller caterpillars would signal less than larger caterpillars, and produce shorter signal units and trains, with lower duty cycles. Results partly supported the hypothesis, showing that in two species there was a positive relationship between size and the number of units produced within two seconds following an attack, the mean number of units per train, and the mean duration of the units in one species. I also tested the hypothesis that sounds of small caterpillars are not in the hearing range of predators. I predicted that there would be a relationship between caterpillar size, and the sound pressure levels and dominant frequencies of the sounds. Results showed no significant relationships with dominant frequencies or sound pressure levels and size. I concluded that the caterpillars made sounds that were within the hearing range of major predators from the onset of sound production. In Chapter Three I followed the other antipredator defences of the four species throughout development. I investigated whether the frequency of defences changed with instar. I found that the caterpillars employed up to seven different secondary defences throughout development. In one species the frequency of dropping and major thrashing increased in the late instars, and in a different species the frequency of regurgitation increased. I concluded that in some cases defence sound production accompanies other secondary defences that increase with the size of caterpillars during development. In Chapter Four I tested the hypothesis that the defensive whistle of the walnut sphinx caterpillar, Amorpha juglandis (Sphingidae: Sphinginae), functions to startle birds. I predicted that the birds would startle to the sounds, and habituate upon repeated exposure within a trial. Results showed that play-back recordings of the whistles elicited a startle response in captive red-winged blackbirds (Agelaius phoeniceus) and caused them to hesitate and/or flee from prey. I concluded that the whistles function as a startle display. Together, the experiments conducted within my thesis addressed important outstanding questions regarding the evolutionary origins of defence sounds in caterpillars, and their functions in predator-prey interactions.