The trophic dynamics of the Yukon boreal forest have been under investigation at the Kluane Lake Research Station since 1973. We monitored and conducted experiments on the major species in this ecosystem, except the large mammals (for logistic reasons). The central problem has been to determine the causes of the 9-10 year cycle of snowshoe hares, and to achieve this we carried out several large-scale experiments manipulating food supplies, predator pressure, and soil nutrient availability to test hypotheses that food, predation, or habitat quality regulate populations. The hare cycle is driven top-down by predators, and most hares die because they are killed by predators. Predators also cause stress in female hares, and the stress response seems to be responsible for the loss of reproductive potential in the decline and low phases of the hare cycle. Many of the specialist predators and some herbivores in this ecosystem fuctuate with the hare cycle. Arctic ground squirrels do, but red squirrels do not, being linked closely to white spruce seed masting years. Small rodents fuctuate in numbers in two patterns. Red-backed voles and four species of Microtus voles have a 3-4 year cycle that seems to be driven by food supplies and social behaviour. Deer mice, in contrast, have fuctuated dramatically in the 38 years we have monitored them, but not cyclically. White spruce seed production varies with temperature and rainfall, but was not affected by adding nutrients in fertilizer. Global warming and reduced hare browsing in the last 20 years have helped to increase the abundance of shrubs in these forests. It will be challenging to predict how this system will change as climatic warming proceeds, because even closely related species in the same trophic level respond differently to perturbations. We recommend continued monitoring of the major species in these boreal forests.