Cephalopods provide forage to a wide range of predators in marine food-webs. Despite their ecological importance, a basic understanding of the mechanisms controlling predation risk and demand is lacking. This is true of one of the most common species of squid found in the northwest Atlantic, the longfin inshore squid (Loligo pealeii). In this dissertation, I address this shortcoming by investigating the role that size and behavior play in influencing squid’s vulnerability to predation. I used long-term food habits, population survey, and commercial landings data, to quantify size-based patterns of predation respective to 25 species of predators. Additionally, I estimated the amount of overlap between predatory consumption and the fishery catch for squid by size. I found that finfish and elasmobranchs generally consumed juvenile and sub-adult squid, while marine mammals primarily targeted adults. Consequently, marine mammals had the highest overlap with the fishing industry for squid size resources. Although large squid were not common in predator diets, predators did not appear to be gape-limited when feeding on squid. This suggested that other factors, including behavior, were important in shaping size-based patterns of predation. I used a laboratory-based approach to quantify attack and capture behaviors towards squid by two predators representing contrasting foraging tactics. Bluefish (Pomatomus saltatrix) and summer flounder (Paralichthys dentatus) were chosen as cruising and ambush predators, respectively. Patterns in attack rates suggested that sizeselection on squid was constrained by passive processes rather than active choice in both predators. Size-dependent profitability functions were calculated by combining capture success rates, handling times, and relative prey mass, and determined that bluefish was the more efficient predator of squid. Lastly, I evaluated the occurrence and effectiveness of anti-predator responses used by squid in the presence of bluefish and flounder. Squid behavior depended on the type of predator present, and the survival value of primary and secondary defense behaviors differed during interactions with each predator. The results of this project are intended to improve the quality of management of squid and their predators by providing a better understanding of predator-prey interactions in the northwest Atlantic.

Em diferentes aplicações de Aprendizado de Máquina podemos estar interessados na minimização do valor esperado de certa função de perda. Para a resolução desse problema, Otimização estocástica e Sample Size Selection têm um papel importante. No presente trabalho se apresentam as análises teóricas de alguns algoritmos destas duas áreas, incluindo algumas variações que consideram redução da variância. Nos exemplos práticos pode-se observar a vantagem do método Stochastic Gradient Descent em relação ao tempo de processamento e memória, mas, considerando precisão da solução obtida juntamente com o custo de minimização, as metodologias de redução da variância obtêm as melhores soluções. Os algoritmos Dynamic Sample Size Gradient e Line Search with variable sample size selection apesar de obter soluções melhores que as de Stochastic Gradient Descent, a desvantagem se encontra no alto custo computacional deles. / In different Machine Learnings applications we can be interest in the minimization of the expected value of some loss function. For the resolution of this problem, Stochastic optimization and Sample size selection has an important role. In the present work, it is shown the theoretical analysis of some algorithms of these two areas, including some variations that considers variance reduction. In the practical examples we can observe the advantage of Stochastic Gradient Descent in relation to the processing time and memory, but considering accuracy of the solution obtained and the cost of minimization, the methodologies of variance reduction has the best solutions. In the algorithms Dynamic Sample Size Gradient and Line Search with variable sample size selection, despite of obtaining better solutions than Stochastic Gradient Descent, the disadvantage lies in their high computational cost.

Em diferentes aplicações de Aprendizado de Máquina podemos estar interessados na minimização do valor esperado de certa função de perda. Para a resolução desse problema, Otimização estocástica e Sample Size Selection têm um papel importante. No presente trabalho se apresentam as análises teóricas de alguns algoritmos destas duas áreas, incluindo algumas variações que consideram redução da variância. Nos exemplos práticos pode-se observar a vantagem do método Stochastic Gradient Descent em relação ao tempo de processamento e memória, mas, considerando precisão da solução obtida juntamente com o custo de minimização, as metodologias de redução da variância obtêm as melhores soluções. Os algoritmos Dynamic Sample Size Gradient e Line Search with variable sample size selection apesar de obter soluções melhores que as de Stochastic Gradient Descent, a desvantagem se encontra no alto custo computacional deles. / In different Machine Learnings applications we can be interest in the minimization of the expected value of some loss function. For the resolution of this problem, Stochastic optimization and Sample size selection has an important role. In the present work, it is shown the theoretical analysis of some algorithms of these two areas, including some variations that considers variance reduction. In the practical examples we can observe the advantage of Stochastic Gradient Descent in relation to the processing time and memory, but considering accuracy of the solution obtained and the cost of minimization, the methodologies of variance reduction has the best solutions. In the algorithms Dynamic Sample Size Gradient and Line Search with variable sample size selection, despite of obtaining better solutions than Stochastic Gradient Descent, the disadvantage lies in their high computational cost.

The aquatic isopod Asellus aquaticus can in some lakes be found as two different ecotypes; one in the habitat dominated by reed and one in stonewort stands. These ecotypes have been shown to differ in size, color and behavior. The reed ecotype is larger, darker and more active compared to the stonewort ecotype. In the two habitats there are different dominating predators: Invertebrate predators in the reed habitat and fish in the stonewort habitat. This project aims to examine how the presence of invertebrate predators and fish affect the two ecotypes of the isopod in regard to behavior and size composition in different substrates. To examine the effect on behavior the activity of isopods collected from Lake Tåkern was measured without and with chemical cues from perch and damselfly larva. The result of the behavior experiment showed no difference between the ecotypes nor the treatments. The lack of differences in the behavior indicates that there could be a variation between lakes. The effect on the size composition was examined by subjecting groups of isopods to predation by perch or damselfly larva in different substrate. The mean length of the group was measured before and after the trials. The size decreased significantly for the reed ecotype in stonewort substrate when subjected to predation by perch. The size for the stonewort ecotype increased significantly in reed substrate with damselfly larva as predator. The effect on size supports that the predators are the cause of the size difference between the ecotypes.

IR-Spektren von neutralen Na(H2O)n-Clustern wurden im OH-Streckschwingungsbereich (2800 – 3800 cm-1) größenselektiv bis n=575±25 gemessen. Die Infrarot-anregungsmodulierte, durch chemische Vorionisation verstärkte Photoionisation beruht auf dem Effekt einer Signalverstärkung durch IR-Absorption bei schwellennaher Photoionisation. Die Strukturinformation kann anhand eines Vergleichs mit quantenchemischen Rechnungen gewonnen werden. Die mit dieser Methode erhaltenen IR-Spektren können für verschiedene Größenbereiche mit durch andere Verfahren generierten Spektren verglichen werden. Für mittlere Clustergrößen (~n=20-50) sind die erhaltenen IR-Signaturen sehr ähnlich zu RIDIR-Spektren von phenoldotierten Clustern. Die Spektren großer amorpher Cluster (n=250) sind vergleichbar mit denen ionischer Cluster. Die Variation der Expansionsbedingungen (z. B. die Änderung des Trägergases) beeinflusst nicht nur das Einsetzten der Kristallisation (beginnend zwischen n=200 und n=400), sondern auch die Isomerenzusammensetzung für kleinere Clustergrößen. Auf dieser Grundlage wurden Strukturen für die Größen (H2O)20, (H2O)25 und (H2O)32 untersucht.

Piscivorous birds are an integrated part of lake and coastline ecosystems. Despite decades of research it is yet unclear what influence fish eating birds have on the fish community. The aim of this literature study was to focus on how two fish eating birds, Great Cormorant and Great Crested Grebe, influence the fish community in shallow and eutrophic lakes. Different types of analysis methods have been used when doing research on the subject, including pellet analysis, stomach content analysis, tagging of fishes by “PIT”-techniques and analysis by observation. Results show that conclusions by studies tend to vary, but generally higher bird density, lower water temperature and a relatively high turbidity contribute to a higher influence on the fish community. Both bird species are opportunistic in their choice of food and catch prey of the species that occur locally. Both species also limit themselves in the prey size, but the choice vary over seasons because of water temperature and the birds requirements in association with breeding and migration. The significant level of influence seems to be when predation is conducted on younger and smaller fishes. The influence on the fish size can also be indirect, where predation on smaller individuals prevent fishes from becoming older and bigger. As studies tend to deviate in conclusion, no general answer to this issue can be given at present. Comprehensive research with several years of full control over both fish-and bird population is needed to find the proper conclusion.