The size of any given population is determined by the birth rate in proportion to the rate of deaths. Factors that will incease the birth rate include a low level of competition for food, mating partners and territory. An increase in population density will cause an increase in abovementioned competition levels. This, for obvious reasons, will have a resultant increase in the death rate. The stress induced by the increased competition will also result in a decrease in birth rate because of depressed oestrus cycles in mammals as well as lesser fertility due to stress. An incease in the secretion of corticoids by the brain during stress is the reason for this. I trust that this will answer your question.

It depends on your definition of "too large", actually, as many populations do get "too large", and then are taken back to a lower level by mass fatalities. If a population outstrips its food source, many starve. As an example, this happens with lemmings and stoats -- lemmings are the main or only food source for stoats in several northern areas. When the lemming population increases, so does the stoat population. However, stoats are also very prolific breeders, and it comes to a point (every four years, in fact) that over-predation strips down the lemming population -- and a lot of stoats starve. Then, in the absence of stoat predation the lemming population begins to build up again and the cycle starts over. Which of course also points out predation as being a population-limiting factor, for the lemmings. See http://abc.net.au/science/news/stories/s979473.htm for a nice summary. Competition for food and resources between different species can also act to reduce what resources are available to each population, even though none of the competing populations get the bulk of the resource. As an example, think of plants competing for sunny spots on a forest floor. There are only so many sunny spots, and they may be claimed on a "first sprouted, first served" basis; subsequent plants don't have the good spots and may not thrive. Additionally, the build-up of certain products in an environment may act as a population limiter; for example, a build-up of waste products from bacterial metabolisms may act to limit further growth of those bacterial populations. The acidification of soil from dropped pine needles can make it more difficult for new trees to sprout. In general, the combination of "resources available [minus] inhibitory factors" is referred to as an environmental carrying capacity. In other situations, disease and/or parasitism is a limiting factor; dense, crowded populations tend to pass on infectious diseases and parasites a lot more easily than sparse, widely separated populations, for what ought to be obvious reasons. An example of density-dependent disease might be myxomatosis in rabbits; an example of parasitism might be beetle infestation in pines -- the more densely plants are crowded together, the more easily a pest invasion can pass from one plant to another. There are other limiting factors to more stable populations, and these often have to do with stress limiting animals' ability to breed. In the 1950s, Edward T. Hall did a survey of animal literature, and ran a number of rat experiments of his own, and demonstrated that overcrowded animals tended to have more hostile and stressful interactions, and that the stress of the environment produced hyperactive adrenal glands -- with a knock-on effect of atrophied gonads and lower fertility. Wolves and some other top social predators actually exploit this effect -- only the top one or two pairs in a pack breed each year, because the stress of continuously getting picked on and knocked around keeps the sex hormones in lower-ranking animals suppressed, and they never become fertile. Another example of stress-limited fertility appears in rabbits, again. Rabbits are capable of "freezing" the development of embryos for up to six months, at which point the pregnancy must either be allowed to continue development or aborted. In overcrowded and stressed conditions, rabbit females may hold onto suspended pregnancies that way for some time, to see if conditions ease and they have sufficient resources to sustain babies; however, if conditions don't ease, then the females will secrete a hormone mix which aborts the pregnancy and either resorbs the embryos for recycling of materials, or (if they are too far along in development) desiccates the embryos and then encysts them. A doe can carry up to 24 encysted embryos, with apparently no ill effects to herself. Some rodents also have similar abilities to self-abort pregnancy under stressful conditions. However, another limiting factor of reproduction, in rodents and many other species, is predation of the young. In overcrowded and stressful conditions, many creatures will try to kill each others' babies (if nothing else, to try to give their own babies a better chance through lessening competition for scarce resources). Other species rarely push their environment's carrying capacity, because rather than a strategy of rapid reproduction of many offspring, they have few offspring and spend a lot of time taking care of them. Organisms that favor rapid, prolific reproduction are called r-strategists; organisms that favor slow, lower-number reproduction with a high degree of parental care are called K-strategists. Each strategy exploits different types of environmental niches; r-strategists tend to have short lifespans, exploit transitory environments, and have good methods of dispersal so that they can spread as widely as possible. K-strategists instead tend to have longer lifespans, exploit stable environments (like mature woodlands), and develop very efficient ways of exploiting narrow niches, and have low infant mortality overall. However, even though their low birth rate means they don't push environmental carrying capacity as much, K-strategists are still subject to the limiting factors above, including predation, disease and starvation. Aside from these density-dependent mechanisms, there are also density-independent events which will have an effect -- droughts, floods, forest fires, particularly hard winters...all these things and other disasters will decrease local populations of many organisms. And there is one final strategy which can help prevent overpopulation -- migration. This is, of course, dependent on there being a suitable environment within reach for the population to spread to, something increasingly rare now because of human dominance of many environments. But traditionally, a response to a certain population density would be for individuals to migrate out of the territory, to an uncolonised territory somewhere. There isn't really a one-size-fits-all answer, but the main themes do tend to be movement, food supply, predation, disease, and control of reproduction, and hopefully the examples provided give you an idea of how it can work.