As formas de defesa encontradas na natureza são incrivelmente diversas e envolvem estratégias que minimizam as chances de encontro com o predador ou as chances de escape diante de um ataque. A liberação de substâncias químicas é uma forma de defesa amplamente difundida entre artrópodes e vários estudos já demonstraram sua eficiência contra o ataque de predadores. Apesar de trazerem benefícios óbvios em termos de defesa, a produção de vários desses compostos químicos deve ser custosa para as presas. Esta tese teve como objetivo geral explorar os custos da produção de defesas químicas em um grupo particular de artrópodes, os opiliões. No capítulo 1, investigamos como a produção de ovos pode comprometer a produção de defesas químicas (benzoquinonas) e quais as conseqüências da redução do volume de secreção liberada sobre a capacidade de defesa das fêmeas de Acutisoma longipes. Nossos resultados apontam que a quantidade de secreção produzida por fêmeas ovígeras é quase 50% inferior à das fêmeas não-ovígeras. Como conseqüência, a secreção liberada por fêmeas ovígeras é menos eficiente em deter formigas e aranhas do que a secreção liberada por fêmeas não-ovígeras. No capítulo 2, investigamos como a quantidade e a qualidade da dieta influenciam a produção de defesas químicas (benzoquinonas) em Magnispina neptunus. Demonstramos que a produção de defesas químicas é condição dependente, pois indivíduos bem alimentados produziram mais secreção que indivíduos mal alimentados. Entretanto, indivíduos que receberam aporte extra dos precursores da secreção não incrementaram nem a quantidade total nem a concentração da secreção. Por fim, o capítulo 3 teve como objetivo investigar possíveis demandas conflitantes (trade-offs) entre o esforço de acasalamento (produção de armamentos) e o esforço somático (defesas químicas) de machos e fêmeas em um clado de opiliões neotropicais. Usando uma abordagem comparativa, mostramos que fêmeas produzem mais defesas químicas (benzoquinonas) que machos, mas não há uma relação negativa entre o dimorfismo sexual morfológico e o dimorfismo na quantidade de defesas químicas produzidas por machos e fêmeas. Coletivamente, os resultados obtidos aqui apontam que a produção de defesas químicas em opiliões é custosa e que está sujeita a demandas conflitantes com outros componentes de aptidão. / The forms of animal defenses found in nature are incredibly diverse and comprise many strategies that minimize the changes of a prey being detected or attacked by a predator. The emission of chemical secretions is a widespread defense among arthropods, and several studies have already demonstrated the efficiency of chemical defenses against predation. Although chemical defenses confer obvious survival benefits, the production of deterrent compounds may be costly for the individuals. The main goal of this thesis was to evaluate the costs of producing chemical defenses in a particular arthropod group, the harvestmen. In chapter 1 we investigated how egg production may compromise the production of chemical defenses (benzoquinones) by females of Acutisoma longipes, and whether a reduction in the amount of secretions released by ovigerous females make then more vulnerable to predation. Our results indicate that ovigerous females produce almost 50% less secretion than non-ovigerous females. Moreover, the low amount of secretion released by ovigerous females is less efficient than the amount released by non−ovigerous females in deterring ants and spiders. In chapter 2 we investigated how the quantity and quality of food influence the production of chemical defenses (benzoquinones) in Magnispina neptunus. We demonstrate that the production of chemical defenses is clearly condition dependent because well fed individuals produced more secretion than poorly fed individuals. However, individuals that received in the diet an input of the benzoquinones\' precursor did not show an increase in the amount and concentration of secretion released. Finally, in chapter 3 we investigated possible trade-offs between mating effort (i.e., investment in weaponry) and somatic effort (i.e., investment in chemical secretions) in males and females of several harvestman species belonging to the family Gonyleptidae. Using a comparative approach, we showed that females consistently produce more secretion than males, but there is no negative relationship between morphological sexual dimorphism and sexual dimorphism in the amount of secretion released by males and females. Taken together, our findings indicate that the production of chemical defenses in harvestmen is costly and is under allocation trade-offs with other fitness components.

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.

<p>White-tailed deer (Odocoileus virginianus; hereafter deer) are a keystone herbivore within forest ecosystems. While deer rely on plant species for growth, reproduction, and survival, multiple external factors can dictate browsing behavior. These factors ultimately drive browsing selection, browsing intensity, and diet composition, which in turn can shape the influence deer have on forest ecosystems. To better understand the complex relationship between deer populations, their habitat, and public perception of deer, the Indiana Department of Natural Resources partnered with Purdue University to initiate the Integrated Deer Management Project (IDMP). As part of the IDMP, this dissertation evaluated the ecological condition of deer habitat to assess the influence deer have on woody and herbaceous plant species within Indiana forests. Our study aimed to i) rank woody species according to their selection by deer and evaluate how the ranking of individual species varies across the state (Chapter 2); ii) evaluate variables and spatial extents associated with differences in browsing intensity, and evaluate different indices used to assess differences in browsing intensity (Chapter 3); iii) quantify winter deer diet composition using DNA barcoding to evaluate how diets vary across a gradient of deer densities, browsing intensities, non-native plant densities, and landscape characteristics (Chapter 4); and iv) evaluate the interactive effects of deer, non-native plant species, and landscape characteristics on the herbaceous layer of forests, while further evaluating the spatial extent at which landscape characteristics are most strongly related to herbaceous-layer composition and diversity (Chapter 5). To do this, I sampled 152 woodlots over three years across three regions of Indiana, collecting data on the browsing selection of individual woody species, the browsing intensities on all woody species, and the composition of vegetation communities (Chapters 2, 3, and 5, respectively). To address diet composition (Chapter 4), we collected deer pellet groups to analyzed diet components. We ranked a total of 63 woody species regarding their browsing selection by deer. While most of these remained consistent from region to region, 16 varied greatly in selection, as deer often showed increased selection for a given species when it resided in an area that provided greater browsing opportunities. Browsing intensity was most associated with food availability, however, it was also influenced by deer density in the region with the lowest forest cover. The twig age index of browsing intensity showed promise as the most efficient and effective index for use in Indiana woodlots. Although diet composition did not differ across regions of Indiana, we found 16 that deer consumed several uncommon taxa when the greater landscape exhibited homogenous patch composition. Similarly, deer consumed different native taxa in forested landscapes with greater deer densities in comparison to agricultural landscapes with lower deer densities. Lastly, though browsing varied within and across regions landscape characteristic, and not deer, were the most influential suite of variables. Additionally, the spatial extent at which these variables exhibited their best fit varied depending on the dependent variable being evaluated and the region of analysis. Together, our results highlight that variables ranging from the woodlot to the landscape-scale influence browsing behavior. This showcases that deer respond to variables at varying scales when browsing and in general, browse more in areas that offer the greatest benefit, whether these areas offer greater food availability or quality, or offer lower risks associated with anthropogenic development. This suggests that when managing forests for deer both woodlot and landscape context should play a role in the decision process. Although differences in browsing were observed, deer had less impact on the herbaceous layer compared to other variables we examined. This suggests that, in contemporary forests, landscape characteristics may be the drivers of changes, and species composition may reflect a long-term history of deer herbivory with less variability resulting from differences in contemporary deer abundance within and across regions.  </p>

Two consecutive experiments were conducted to test a hypothesis that mixed grazing outcome is influenced by the type of stocking system applied. The objective of both experiments was to investigate the influence of co-grazing with sheep on cattle liveweight gain (LWG) under continuous (C) and rotational (R) stocking, where sheep weekly liveweight change under the two stocking systems was kept similar. In experiment I nine yearling heifers (266 ± 4.5 kg liveweight) and 27 ewe hoggets (54±0.9 kg liveweight) were continuously stocked for 19 weeks on an irrigated perennial ryegrass-white clover pasture (2.95 ha) maintained at a sward surface height (SSH) of 5cm by adding or removing additional animals in a fixed ratio (1: 1 W⁰.⁷⁵ cattle:sheep). An equal area of pasture was rotationally stocked by a similar group of animals where they received a new area of pasture daily and also had access to the grazed area over the previous 2 days. The size of the new area provided daily was such that the weekly liveweight change of rotationally co-grazed sheep was equal to that of those continuously co-grazed with cattle. Similar groups of animals were used in the second experiment with additional group of 9 heifers grazed alone on C and R pastures. Liveweight of animals was recorded weekly and final fasted weight was determined after 24-hour total feed restriction. SSH on both treatment swards was recorded daily. There were three intake measurement periods spread over the trial period. Organic matter intake (OMI) was predicted from the ratio of N-alkanes in faeces and herbage. Diet composition was determined by dissecting oesophageal extrusa samples. Grazing behaviour (bite rates and grazing time) were also recorded. The mean SSH for C pasture was 5.1±0.09 cm. Overall pre- and post-grazing SSH for R pasture was 15.9 ±0.12 and 5.6 ±0.07 cm, respectively. As determined by the protocol average daily LWG of sheep was similar between C and R (147 (±5.8) vs 138 (±6.7) g day⁻¹; (P>0.05). In contrast, cattle continuously stocked with sheep grew 200 g day⁻¹ slower than those rotationally stocked with sheep (800 (±41.6) vs 1040 (±47.7) g day⁻¹, P<0.0l). R heifers achieved 30 kg higher final fasted liveweight than C heifers (350 vs 381 kg; P<0.01). Overall LWG per ha was also 6 % higher under R than C stocking (674 vs 634 kg ha⁻¹). The OMD of both sheep (73.5 vs 75.8 %) and cattle (75.8 vs 78.0 %) diets was similar under continuous and rotational stocking. There was no significant difference OMI data also concurred with the L WG data (Cattle: 7.94 vs 6.31 (±0.32) kg day⁻¹ (P<0.05); sheep: 1.40 vs 1.44 (±0.04) kg day⁻¹ for Rand C treatments, respectively). There was no difference in clover content of cattle diet under C and R treatments. C heifers had higher number of bites per minute than R heifers (62 vs 56; P<0.05). Proportion of heifers seen grazing (every 15-minute) during four 24-hour observations was greater on C than R pasture (0.44 vs 0.31 (±0.03); P<0.05). The similarity coefficient between sheep and cattle diet was 0.61 and 0.76 under C and R stocking, respectively. The lower daily LWG of C heifers was attributed to (a) the lower SSH under C than R stocking and/or (b) the inability of cattle to compete well with sheep where there is small, continual renewal of resources (C) in contrast to a large periodic renewal under R stocking. This experiment showed that the outcome of mixed gruing can be influenced by the stocking system chosen. But it was not possible to apportion the difference in LWG of cattle between mixed grazing per se and the difference in mean grazed sward height (5.1 for C vs 10.8 cm for R). A second experiment was conducted to determine the relative performance of cattle co-grazed with sheep (CS) and grazed alone (CA) under each stocking system. Hence, there were four treatments. CA- continuous stocking (CA-C), CS- continuous stocking (CS-C), CA- rotational stocking (CAR) and CS- rotational stocking (CS-R). A total area of 4.42 ha was allocated to each stocking system. Under C stocking, 2.95 ha (2/3) was assigned to CS-C and 1.47 ha (1/3) to CA-C, and SSH on both treatments was kept at 4 cm by adding or removing extra animals. Under R stocking, CA-R and CS-R grazed side by side separated by an electric fence. They were given a fresh area daily, the size of which was varied such that the weekly LW change of R sheep was equal to that of the C sheep. CA-R received one-third of the new area though the size was adjusted regularly to achieve the same post-grazing SSH with CS-R. Measurements included: weekly liveweight change, OMI (two periods) and diet composition (using N-alkanes). The mean SSH of CA-C and CS-C swards was 4.27 and 4.26 (±0.02) cm, respectively. CA-R and CS-R swards had mean pre-grazing SSH of 14.9 and 15.2 (±0.08) cm and post-grazing heights of 4.87 and 4.82 cm (±0.03), respectively. The proportion of areas infrequently grazed was higher for CA-C than CS-C swards (0.22 vs 0.17, respectively). C and R sheep daily LWG: 155 (±0.6) and 147 (±0.7) g, and OMI: 1.96 and 2.04 (±0.ll) kg, respectively, were not significantly different. They also had similar diet composition. In comparison, CS-C heifers grew only at 69 % of the daily LWG achieved by CS-R heifers (706 vs 1028 (±72) g; P<0.05). LWG of CA-C and CA-R was 916 and 1022 (±72) g day⁻¹, respectively. The difference in LWG between CS-R and CS-C (D₁) heifers was due to difference in mean sward height, stocking system and mixed grazing, while D₂ (difference in LWG between CA-R and CAC) was due to difference in mean sward height and stocking system. D₁-D₂ (the effect of stocking system on mixed grazing) was 216 g and made up 67 % of the total difference between CS-R and CS-C. There was a significant stocking system-species mixture interaction in the final fasted LW achieved by heifers. Final fasted LW was significantly lower for CS-C than CA-C heifers (283 vs 323 (±9.7) kg), but did not differ between CS-R and CA-R (332 vs 330 (±9.7) kg, respectively). The digestibility of diet OM was similar for both continuously and rotationally stocked sheep (84.4 vs 83.2 %, respectively). Cattle diet OMO was 76.5, 74.7, 79.4 and 77.8 for CA-C, CS-C, CA-R and CS-R respectively (P>0.05). Differences in OMI followed a similar pattern to daily LWG. Mean daily OMI was 8.98, 6.24, 8.80 and 9.45 (±0.40) kg for CA-C, CS-C, CA-R and CS-R, respectively. Clover content of the diet of CA-C heifers was three times higher than that of CS-C heifers (30.7 vs 10.4 % OM; P<0.05); there was no difference in clover content of diets of CS-R and CA-R heifers (21.5 vs 23.9 % OM, respectively). In both stocking systems LWG per ha was higher on CA than CS treatments. These results suggested that the disadvantage of selective clover grazing by sheep outweighed the advantages of sheep grazing around cattle dung patches under continuous stocking. Under rotational stocking, rapid diurnal changes in sward conditions probably limited selective grazing by both sheep and cattle such that there was no disadvantage to CS cattle. The results do not provide a basis for recommending grazing cattle with sheep rather than cattle alone, but do provide some basis for recommending co-grazing of sheep and cattle using rotational rather than continuous stocking.