Back in 2001, the paradoxically high biodiversity on Barro Colorado Island inspired Hubbell to propose the groundbreaking neutral theory of ecology. Traditional ecology theory stressed the competition for niches between species. But Hubbell pointed out that species might not really matter in that equation because, in effect, individuals compete for resources with members of their own species too. He suggested that patterns of diversity in ecosystems might largely be the products of random processes.

For a theory that dealt with biodiversity, Hubbell’s neutral theory was sparse. It ignored variations in life spans, nutritional quirks and other details that distinguish one species from another. In models based on the theory, every individual in a theoretical ecosystem is identical. Once the clock starts, the ecosystem evolves stochastically, with individuals outcompeting and replacing each other at random. The theory was completely at odds with species-based approaches to ecology, and it provoked impassioned debate among ecologists because it seemed so counterintuitive.

Yet surprisingly, as the random walks in the neutral models progressed, they reproduced key features of what Hubbell and his colleagues saw in their data from Barro Colorado Island and what others have seen elsewhere. In this modeling that almost perversely acknowledges no differences, there are flashes of the real world.

That tension between the models and reality has long interested O’Dwyer. Why did neutral theory seem to work so well? Was there a way to bring in information about how species function to get results that might look still more realistic?

One of the things that make neutral models appealing, O’Dwyer said, is that there really are deep universalities among many living things. While animal species are not identical, they are remarkably similar at the level of, say, the circulatory system. The same numbers concerning physiology crop up again and again in animals and plants, reflecting perhaps the constraints of their shared evolutionary history. According to a principle called Kleiber’s law, for example, the metabolic rate of an animal generally increases with its size, scaling as a power law — the same power law, no matter the species. (Several theories about why Kleiber’s law is true have been offered, but the answer is still debated.)

Given those signs of underlying order, O’Dwyer wondered whether some details of how organisms live matter more than others in determining how successfully species will compete and survive over evolutionary time. Take metabolism again: If an ecosystem can be seen as an expression of its inhabitants’ metabolisms, then the organisms’ sizes are special, significant numbers. The size of an individual may be more useful in modeling its fate over time than any number of other details about its diet or species identity.

O’Dwyer wondered whether one of those crucial, privileged factors might be captured by life history, a concept that combines species statistics such as average number of offspring, time until sexual maturity and life span. Imagine a plot of 50 individual plants. Each has its own life span, its own pattern of reproduction. After three months, one plant might produce 100 seeds, while another similar one produces 88. Maybe 80% of their seeds will germinate, producing the next generation, which will go through its own version of this cycle. Even within a species, individual plants’ numbers will vary, sometimes by a little, sometimes by a lot, a phenomenon called demographic noise. If this variation is random, in the manner of Hubbell’s neutral theory, what patterns will emerge over successive generations?

O’Dwyer knew he had found someone who could help him explore that question when Jops joined his lab as a graduate student. Jops had previously studied whether models using life histories could predict whether a vulnerable plant species would survive or if it was on the way out. Together, they started to hammer out the math that would describe what happens when life history meets competition.

In Jops and O’Dwyer’s model, as in neutral models, stochasticity — the influence of random factors on deterministic interactions among the species — is important. The life histories of species, however, can amplify or reduce the effects of that randomness. “Life history is a kind of lens through which demographic noise works,” O’Dwyer said.

When the researchers allowed their model to progress through time, putting each simulated individual through its paces, they found that certain species could persist alongside each other for long periods even though they were competing for the same resources. Looking deeper into the numbers for an explanation, Jops and O’Dwyer found that a complex term called effective population size seemed useful for describing a kind of complementarity that could exist among species. It encapsulated the fact that a species could have high mortality at one point in its life cycle, then low mortality at another, while a complementary species might have low mortality at the first point and high mortality at the second. The more similar this term was for two species, the more likely it was that a pair could live alongside each other despite competing for space and nutrition.

“They experience demographic noise at the same amplitude,” O’Dwyer said. “That’s the key for them to live together a long time.”