Nature is full of alliances, mutualistic partnerships between two species. Flowers offer bees and butterflies nectar in exchange for pollination. Anemones offer clown fish shelter in exchange for protection.

But new research suggests these relationships are highly dependent on the environment. And like the environment, these relationships aren't permanent.

A team of scientists from the Massachusetts Institute of Technology and Harvard University studied the relationship between two mutualistic strains of yeast. Because each of the two yeast strains produces an amino acid useful to the other, they both benefit by cross-feeding.

To see how environmental factors affected their relationship, researchers gradually increased levels of available nutrients. Adding nutrients made the environment increasingly "benign."

When their environment featured fewer nutrients, the

In low nutrient conditions, both strains thrived when together and struggled on their own. With a small amount of added nutrients, the strains were still obliged to cooperate, a relationship researchers called "obligate mutualism." When scientists made nutrients slightly more plentiful, the yeasts survived on their own, but still did better together. They dubbed the relationship "facultative mutualism."

When researchers flooded the environment with nutrients, the two strains became competitors, with one strain thriving and the other losing out. This was "parasitism." When even more nutrients were added, the relationship plummeted further, with even the winner struggling and the loser driven to extinction.

"What's amazing is, often when we talk about these interactions between species, we say, 'Oh, a clownfish and an anemone is mutualism, whereas a lion and an antelope is predator-prey.' We talk about these species having fixed interactions," lead researcher Jeff Gore, a professor of physics at MIT, said in a news release. "Whereas here we see these strains go through all these different regimes, just by changing one knob."

Using their observations, Gore and his colleagues created a model to predict the state of the relationship between two strains based on the growth and evolution of the strains over time.

While the findings may not translate perfectly to the study of relationships between coral and algae or ants and acacia trees, researchers say the findings -- detailed in the journal PLOS Biology -- suggest mutualism is most likely to thrive in stressful environs.

"There's a general idea that more challenging environments favor mutualistic type interactions," Gore said. "These experiments provide further support for the idea that mutualisms will often break down or become more competitive in more benign environments. Which is something that people have seen some evidence for in natural populations, but this is a nice context in which we can see it happening in a very direct way."