At first glance the story of the Intense Pulsed Neutron Source (IPNS) at Argonne National Laboratory (ANL) appears to have followed a puzzling course. When researchers first proposed their ideas for an accelerator-driven neutron source for exploring the structure of materials through neutron scattering, the project seemed so promising that both Argonne managers and officials at the laboratory's funding agency, the Department of Energy (DOE), suggested that it be made larger and more expensive. But then, even though prototype building, testing, and initial construction went well a group of prominent DOE reviewers recommended in fall 1980 that it be killed, just months before it had been slated to begin operation, and DOE promptly accepted the recommendation. In response, Argonne's leadership declared the project was the laboratory's top priority and rallied to save it. In late 1982, thanks to another review panel led by the same scientist who had chaired the panel that had delivered the death sentence, the project was granted a reprieve. However, by the late 1980s, the IPNS was no longer top priority within the international materials science community, at Argonne, or within the DOE budget because prospects for another, larger materials science accelerator emerged. At just this point, the facility started to produce exciting scientific results. For the next two decades, the IPNS, its research, and its experts became valued resources at Argonne, within the U.S. national laboratory system, and within the international materials science community. Why did this Argonne project prosper and then almost suffer premature death, even though it promised (and later delivered) good science? How was it saved and how did it go on to have a long, prosperous life for more than a quarter of a century? In particular, what did an expert assessment of the quality of IPNS science have to do with its fate? Getting answers to such questions is important. The U.S. government spends a lot of money to produce science and technology at multipurpose laboratories like Argonne. For example, in the mid-1990s, about the time the IPNS's fortunes were secured, DOE spent more than $6 billion a year to fund nine such facilities, with Argonne's share totaling $500 million. And an important justification for funding these expensive laboratories is that they operate expensive but powerful scientific tools like the IPNS, generally considered too large to be built and managed by universities. Clearly, 'life and death' decision making has a lot to tell us about how the considerable U.S. federal investment in science and technology at national laboratories is actually transacted and, indeed, how a path is cleared or blocked for good science to be produced. Because forces within Argonne, DOE, and the materials science community obviously dictated the changing fortunes of the IPNS, it makes sense to probe the interactions binding these three environments for an understanding of how the IPNS was threatened and how it survived. In other words, sorting out what happened requires analyzing the system that includes all three environments. In an attempt to find a better way to understand its twists and turns, I will view the life-and-death IPNS story through the lens of an ecological metaphor. Employing the ideas and terms that ecologists use to describe what happens in a system of shared resources, that is, an ecosystem, I will describe the IPNS as an organism that vied with competitors for resources to find a niche in the interrelated environments of Argonne, DOE, and the materials science community. I will start with an explanation of the Argonne 'ecosystem' before the advent of the IPNS and then describe how the project struggled to emerge in the 1970s, how it scratched its way to a fragile niche in the early 1980s, and how it adapted and matured through the turn of the 21st century. The paper will conclude with a summary of what the ecosystem perspective shows about the life and death struggle of the IPNS and reflect on what that perspective reveals about how research is produced in the laboratory.