I characterized dusky dolphin (Lagenorhynchus obscurus) underwater bait-balling behaviors and acoustic signals, and compared data between Argentina and New Zealand (NZ) to investigate the roles of ecology versus social learning. I quantified prey herding and capturing behaviors from video footage, and I analyzed acoustic signals from narrowband recordings. In both locations, I related bait-balling behaviors and acoustic signals to group and prey ball sizes. In NZ, I also related dolphin behaviors to prey ball escape behaviors and acoustic signal parameters to examine proximate functions. Observed herding behaviors typically involved dolphins swimming around or under a prey ball using a side body orientation, while dolphins typically captured fish from the side of a prey ball using a ventral orientation. Coordinated prey-capture behaviors may have made it easier for dolphins to capture fish by trapping fish between dolphins. Signals were categorized as click trains, burst pulses, and combinations due to a bimodal inter-click interval distribution. I observed 3 whistle-like chirp-screams, but no whistles. Sequences of burst pulses also occurred that contained 2-14 burst pulses that aurally and visually appeared closely matched. Similarities between locations suggest that ecological context related to broad behavioral and acoustic parameters, while social learning differences may occur on a finer scale. In NZ, prey balls exhibited horizontal and vertical movements, but the only behavior that preceded escape was “funneling”, the brief formation of a ball shape where the height was at least twice the width. Dolphin behaviors that related to prey balls ascending were type of herding pass, location of prey-capture attempts, and body orientation during attempts. These behavioral parameters may also be used to counter vertical prey escape behaviors. In NZ, all signal categories had a direct or indirect role in capturing prey. Click train-burst pulses were likely used for echolocating on prey, burst pulses and sequences appeared to have communication roles, and the role of click trains was ambiguous. No signal categories appeared to have a herding function, but the sheer number of signals emitted may have caused fish to cluster together more tightly and therefore facilitated capture.

Made available in DSpace on 2016-04-29T13:17:49Z (GMT). No. of bitstreams: 1 Paulo Andre Barbosa Panetta.pdf: 927797 bytes, checksum: 1ff8351aa10bad0896b16ccf3e937863 (MD5) Previous issue date: 2013-07-01 / The production of delay of reinforcement with abstains of responses in the delay was studied with four rats. Four experiments were conducted, with one rat in each experiment, and a unique procedure for each study. It was planned an increase in one second upon the duration of the delay in Experiment I, after every three consecutive sessions without responding during the delays. In Experiment II, the delay of reinforcement was kept constant at eight seconds. Experiment III was identical to the previous one, except that the delay periods were signaled. The delay was also signaled in Experiment IV, with an increase of one second on it s duration after every three consecutive sessions without responding during the delay periods. It was only in the last experiment that the occurrence of three consecutive sessions without emissions on the delay periods was observed. This happed on Phases I and II, but not in Phase III. In some few sessions of Experiment I, there was no responding during reinforcement delay, but not for three consecutive sessions. In Experiments II and III, there was not a session without responding upon the delays periods. But it was recorded a reduction in the frequency of emissions during the sessions of both of the experiments, especially in Experiment II. The results indicate that, unless there is an explicit training, emissions will be recorded in the delay periods. The results were discussed from the procedures outlined for each experiment, focusing on the signal functions, the length of delay and the different effects between reinforcement delay resetting and non resetting / Quatro ratos participaram do estudo, que tinha como objetivo geral produzir períodos de atraso fixo de reforço non resetting, através de um treino com resetting, aumento gradual na duração do período de atraso e sinalização apresentada durante o período de atraso. O estudo foi dividido em quatro experimentos, com a participação exclusiva de um sujeito em cada. Foi delineado um procedimento distinto em cada experimento. No Experimento I, foi planejado um aumento de um segundo na duração do atraso de reforço a cada três sessões consecutivas sem emissões nos períodos de atraso. No Experimento II, a duração do atraso de reforço foi mantida constante em oito segundos, sem aumento gradual. O Experimento III foi idêntico ao anterior, com a diferença que os períodos de atraso eram sinalizados pela apresentação de som. No Experimento IV, foi planejado um aumento de um segundo na duração do atraso de reforço, sinalizado pela apresentação do som, a cada três sessões consecutivas sem emissões nos períodos de atraso de reforço. Nos resultados, foi observado apenas no ultimo experimento a ocorrência de três sessões consecutivas sem emissões nos períodos de atraso, durante as Fases Experimentais I e II. Porém, não na Fase Experimental III. Em algumas sessões do Experimento I, não ocorreram emissões durante o atraso de reforço, mas não por três sessões consecutivas. Nos Experimentos II e III, não houve o registro de uma sessão sem emissões nos períodos de atraso, mas foi registrado um redução na frequência de emissões ao longo das sessões dos experimentos, principalmente no Experimento II. Os resultados apontam que, caso não ocorra um treino explicito, emissões serão registradas nos períodos de atraso. Os resultados foram discutidos a partir dos procedimentos delineados para cada experimento, enfocando as funções da sinalização, a duração dos períodos de atraso e os efeitos distintos entre atraso de reforço non resetting e resetting