Over the past decade, the banking of wild seeds has converged with breeding science and agricultural practices, propelled by emerging concerns around the cataloging, conserving, and use of crop wild relatives. The ritually repeated statement that “crop wild relatives have the potential to significantly contribute to climate change adaptation”—by providing domesticates with novel resistances to drought, salinity, or pests—forms the basis for a new, transnational project of institutional collection and breeding. Domesticated plants must be saved by their wild cousins, and agricultural growth with them—but with what effects for wild seed banks, for the relational becomings of wild and domesticated plants, and for the regimes of knowledgeable care enacted by breeders and botanists?

To earn the wild relative descriptor, a plant must be both kin to a domesticated crop species and capable of entering an agricultural relation of “prebreeding,” or experimental crossing with domesticates. Crop wild relatives are not newcomers: they have been pulled into the evolutionary history of wine grapes, to combat phylloxera infestations in the 1960s; of potatoes, to mitigate the potato blight that ravaged Europe in the nineteenth century; and of sugar cane, for crop improvement in the early twentieth century. Increasing scientific specialization has, however, led to applied plant breeding and botanical science being managed separately, with the former mostly conducted at gene banks and guided by food and agriculture policy and the latter confined to botanical gardens and directed toward environmental targets set by international biodiversity conventions (Eastwood et al. 2015). The Crop Wild Relatives Project, launched in 2011 and managed by the Crop Trust in partnership with the Millennium Seed Bank (Dempewolf et al. 2014), represents an attempt to close this gap and enable a reconvergence of agricultural and conservationist concerns—with powerful transformative effects on conservation practices, whose priority and target species may now be reevaluated through the lens of their proximity to domesticated cousins.

The Crop Wild Relatives Project is inscribed into a long history of colonial botanical management, and enrolls conservation biology into an agenda of enclosure, value extraction, and the production of centralized, commodifiable biological resources (Montenegro de Wit 2016). At the same time, the material practices emerging around crop wild relative conservation also destabilize entrenched definitions of “wild” and “domesticated,” forcing renewed attention to (and care for) the marginal spaces in which crop wild relatives evolve and proliferate. At the Millennium Seed Bank and international partner institutions, scientists bring together two categories of subjects that must be made to cooperate across the divide that was created when the globalized logic of agribusiness severed the rich ties between cultivated plants and the wild cousins proliferating at the margins of their plots. Cultivated crops must borrow the resilience of plants that have tested their mettle undisturbed against a variety of environmental hardships; wild relatives must be brought into the securitized spaces of agricultural management, where they will enjoy protection against extinction. In focusing on the retrieval of resilient traits and modes across categories, scientists are paying attention to what happens between beings and species, exercising power not merely on forms of life but through the relations between them.

Crop wild relative conservation and prebreeding thus take place in a variety of borderlands, in particular the weedy margins where crop wild relatives have evolved. As scientists tend to these margins—connecting agriculture to wild relatives, choreographing new interspecies relations in experimental settings—they are called to develop new forms of care, attention, and knowledge. Wild relatives are, in fact, taking botanists and curators on unexpected excursions: before even dreaming of becoming farmers, it seems, they must first turn themselves (again) into hunters. Even wide-ranging efforts such as the Crop Wild Relatives Project yield, at best, uncertain archives, for instance when recalcitrant species such as wild bananas refuse to germinate ex situ in sufficient quantities for prebreeding. This sends breeders back into the field, enrolling them in short-lived but intense attunements to the narrow ecological niches in which plants with specific resistances might be found—“hunting,” as the uncritical extractivist narrative goes, for new forms of drought tolerance.

Crop wild relatives also reveal entrenched gaps in Western scientific knowledge about the accidents and variegation of plant life, forcing scientists to conduct quasi-forensic work in order to understand past interspecies interactions that have shaped beets—or to recover lost knowledge about the reproductive abilities of ensets, by piecing together remainders of erased local cultivation practices and detailed field observations (Tamrat et al. 2020). Collecting and prebreeding thus emerge as sites of difficult interspecies achievement and as a potential proto-communal space of mutual influence, in which the de novo domestication of wild relatives is never a guaranteed outcome and where biologists must constantly re-enact the very first steps of taming and reciprocal capture.

The cosmology sketched out in crop wild relative conservation projects is one in which “domestication” operates both as an ordering device and as a contested category, to be questioned and renegotiated. Crop wild relative conservation cannot sustain the expansion of human agency into the weedy margins for very long: its vitality as a project depends on the continued ability of wild relatives to evolve on their own and to develop new strategies in the face of environmental changes. Therein lies the constitutive paradox of crop wild relative conservation: should the spaces in which wild and domesticated relatives meet be swallowed up entirely by domestication, integrated into the sameness of agro-industrial landscapes, the benefits that could be reaped from wild relatives will eventually dry up and disappear. If breeders and botanists want to care for lineages—for abilities, for the possibility of plastic reactivity to environmental disturbance, for the ongoing exchange of traits across domestication divides—they must engage in practices of fostering the margins in which crop wild relatives proliferate, destabilizing the hegemonic projects and narratives of domestication even as they also reprise them.

Even as it engages in violent acts of extraction and management, crop wild relative conservation highlights how contemporary processes of domestication proliferate in scientific settings, and how crucial it is to name these processes as such if we are to grasp how they may enable or foreclose the possibility of responsiveness and mutual transformation (Despret 2004), accept or refuse to compose with complexly interdependent wild agents (Morizot 2018), and entrench or denaturalize political and ethical relations based on the continued production of wild frontiers and heroic domestication narratives.