Conservation biologists have long been concerned that species extinction could have significant consequences for the stability of entire ecosystems—groups of interacting organisms and the physical environment that they inhabit. An ecosystem could survive the loss of some species, but if enough species were lost, the ecosystem would be severely degraded. In fact, it is possible that the loss of a single important species could start a cascade of extinctions that might dramatically change an entire ecosystem. A good illustration of this occurred after sea otters were eliminated from some Pacific kelp (seaweed) bed ecosystems: the kelp beds were practically obliterated too because in the absence of sea otter predation, sea urchin populations exploded and consumed most of the kelp and other macroalgae.

It is usually claimed that species-rich ecosystems tend to be more stable than species-poor ecosystems. Three mechanisms by which higher diversity increases ecosystem stability have been proposed. First, if there are more species in an ecosystem, then its food web will be more complex, with greater redundancy among species in terms of their nutritional roles. In other words, in a rich system if a species is lost, there is a good chance that other species will take over its function as prey, predator, producer, decomposer, or whatever role it played. Second, diverse ecosystems may be less likely to be invaded by new species, notably exotics (foreign species living outside their native range), that would disrupt the ecosystem’s structure and function. Third, in a species-rich ecosystem, diseases may spread more slowly because most species will be relatively less abundant, thus increasing the average distance between individuals of the same species and hampering disease transmission among individuals.

Scientific evidence to illuminate these ideas has been slow in coming, and many shadows remain. One of the first studies to provide data supporting a relationship between diversity and stability examined how grassland plants responded to a drought. Researchers D. Tilman and J A. Downing used the ratio of above-ground biomass in 1988 (after two years of drought) to that in 1986 (predrought) in 207 plots in a grassland field in the Cedar Creek Natural History Area in Minnesota as an index of ecosystem response to disruption by drought. In an experiment that began in 1982, they compared these values with the number of plant species in each plot and discovered that the plots with a greater number of plant species experienced a less dramatic reduction in biomass. Plots with more than ten species had about half as much biomass in 1988 as in 1986, whereas those with fewer than five species only produced roughly one-eighth as much biomass after the two-year drought. Apparently, species-rich plots were likely to contain some drought-resistant plant species that grew better in drought years, compensating for the poor growth of less-tolerant species.

To put this result in more general terms, a species-rich ecosystem may be more stable because it is more likely to have species with a wide array of responses to variable conditions such as droughts. Furthermore, a species-rich ecosystem is more likely to have species with similar ecological functions, so that if a species is lost from an ecosystem, another species, probably a competitor, is likely to flourish and occupy its functional role. Both of these, variability in responses and functional redundancy, could be thought of as insurance against disturbances.

The Minnesota grassland research has been widely accepted as strong evidence for the diversity- stability theory; however, its findings have been questioned, and similar studies on other ecosystems have not always found a positive relationship between diversity and stability. Clearly, this is a complex issue that requires further field research with a broad spectrum of ecosystems and species: grassland plants and computer models will only take us so far. In the end, despite insightful attempts to detect some general patterns, we may find it very difficult to reduce this topic to a simple, universal truth.

一直以来，生物保护学家一直关注的一个问题是，物种灭绝可能会对整个生态系统的稳定性产生严重的影响，影响到与它们有交集的生物以及它们所生活的物理环境。一个生态系统可以承受某些物种的损失，但如果损失的物种足够多，生态系统就会严重退化。事实上，一个重要的物种损失就可能引发一系列的物种灭绝，可能极大地改变整个生态系统。海獭被从太平洋海藻（海草）床生态系统中移除就是一个很好的例子：没有了海獭捕食，海胆数量暴增，吃掉了大量海带和其他大型海藻，海草床也因此不复存在。 一般来说，物种丰富的生态系统往往比物种贫乏的生态系统更稳定。有人提出，物种的多样性透过三种机制增加了生态系统的稳定性。首先，一个生态系统的物种越多，它的食物网就越复杂，物种间的营养作用方面就会有更多的重合。换句话说，在一个物种丰富的系统里，如果失去了一个物种，就给了其它物种一个很好的机会，代替它扮演猎物、捕食者、生产者、分解者、或任何它所扮演的角色。第二，多样的生态系统不太可能遭到新物种的入侵，特别是会破坏生态系统结构和功能的外来物种（生活在本土范围以外的物种）。第三，在一个物种丰富的生态系统中，疾病可能会传播得比较慢，因为每个物种的数量相对较少，从而增加了同一物种内个体间的平均距离，从而阻碍了个体间的疾病传播。 很多现象我们仍无法解释，但是已经慢慢有一些科学证据来证明这些观点。最早提供数据来支撑多样性和稳定性之间的关系的是一项关于草原植物如何对抗干旱的研究。D. 蒂尔曼和J A.唐宁两位研究员在美国明尼苏达州的雪松溪自然历史保护区，用207个地块在1988（经过两年的干旱）和1986（干旱前）的地表生物量的比值来反映生态系统受干旱破坏的程度。在一个1982年开始的实验中，他们比较了这些比值与每个取样点的植物物种数，发现取样点的植物物种越多，其生物数量在干旱中急剧下降得就越少。那些物种数量超过十种的取样点在1988年的生物量大概是1986年的一半，而那些物种数量少于五种的取样点在经过两年干旱之后，大概只剩下八分之一的生物数量。显然，物种丰富的地块很可能包含一些抗旱植物品种，这些品种在干旱年份生长得更好，弥补了那些不那么抗旱的品种的数量。 从更一般的意义上来说，一个物种丰富的生态系统可能更稳定，因为它更可能含有能对某种气候条件（如干旱）的做出各种反应的物种。此外，物种丰富的生态系统更可能包含有着类似生态系统的功能，因此，如果生态系统中的一个物种损失，另一个物种，可能是竞争对手，就有希望蓬勃发展，并取代它的作用。对恶劣条件的不同的反应，以及系统内的功能重合性，这两个特征都是对抗恶劣气候的保障。 明尼苏达草原的研究作为多样性-稳定性理论的有力证据已被广泛接受，但是，它的研究结果已经被质疑。另外，其他生态系统的类似的研究并没有找到多样性和稳定性之间积极的相关性。显然，这是一个复杂的问题，需要进一步的关于更多各种生态系统和物种的实地研究，对草地的研究和建立计算机模型只能证明部分问题。最后，尽管这些富有洞察力的尝试得出了一些概括性结论，我们可能会发现很难将这个话题缩小成一个简单的、普遍认可的事实。