more than 200 species without recovery plans, more than half had been listed for
three or more years (GAO 1992). The recovery of species under these circumstances
is one of the greatest challenges to the application of ecological science.

In addition to being delayed, recovery plans often have weak goals. A review of the
314 approved recovery plans for threatened and endangered species that were
approved by the U.S. Fish and Wildlife Service and the National Marine Fisheries
Service as of mid 1991, found that population goals were often no higher than existing
population densities at the time of listing (Tear et al. 1993). More than half of the
vertebrates would remain in serious risk of extinction even if they met the population
targets in their recovery plans. In some cases, habitat destruction was so severe that
the recovery plans had little chance of success. The reviewers concluded that,
"Recovery plans all too often "manage for extinction" rather than for survival" (Tear et
al. 1993).

The ultimate goal of the Endangered Species Act is to restore populations so that they no longer are threatened with extinction. When that state is reached, the Act provides for delisting of the species.

III. THE ROLE OF SCIENCE IN THE ENDANGERED SPECIES ACT

Scientific information is needed for implementing all of the processes specified in the Endangered Species Act. The more high quality science is used, the more effectively and more efficiently the Act can achieve the important goals society has asked it to accomplish.

A. Use of Science in the Listing Process

Listing a species as threatened or endangered is the first step in conferring legal protection. It is the conclusion to a decision-making process that draws heavily on ecological science, particularly for assessing the level of risk to a species and developing priorities for listing.

Species are proposed for protection because they are thought to be in danger of extinction or at risk of becoming endangered with extinction. For species deserving protection, delaying the decision to provide protection and recovery will bring most of these vulnerable species even closer to the brink of extinction, restrict the options available for achieving recovery, and increase the eventual cost of the recovery

5

process. Therefore, streamlining the listing process can increase the effectiveness of the Act in achieving its goals and potentially reduce the total costs of doing so.

There is no scientific reason why listing, which is an administrative decision based on the available information, should require much time or agency resources. The uncertainty that may result from sparse information is part of the risk that is evaluated during the listing process. Adding independent peer review or other administrative processes to the listing process would unnecessarily lengthen the time to make a listing decision without providing any substantial benefits. The major problem with the listing process has been its slowness, not inadequacy of the quality of the listing decisions.

1. Which Biological Units Should be Listed?

In the language of the Act, a "species" is taken to include any subspecies of fish or
wildlife (including invertebrates such as insects, crustaceans, and mollusks) or plant
(including fungi). For vertebrates, any distinct population segment of a species, that is
one with unique morphological features or genetic traits, qualifies as a species. How
distinct is distinct enough must be judged on a case-by-case basis. The meaning of
"species" in the language of the Act is, therefore, somewhat imprecise, but the
wording recognizes that a species is made up of an assemblage of individuals that
collectively express genetic, morphological, and behavioral variation, and that this
variation is the basis of evolutionary change and adaptation.

The scientific justification for extending protection to distinct population segments of species is that genetic diversity provides the raw material for adaptation of a species to changing conditions. A wide geographic range decreases the likelihood that a catastrophic event such as wildfire, disease, or alien species introduction could wipe out an entire species. The capacity to respond to environmental change through ecological and evolutionary processes is enhanced by large population size, extended geographical distribution (including spatial structure among its populations), and intraspecific genetic diversity. Therefore, because loss of specific population segments can contribute to the decline of a population and increase the probability of its extinction, protection of population segments is biologically appropriate.

The National Marine Fisheries Service has introduced the concept of an "evolutionarily significant unit" to better define and identify distinct population segments. An evolutionarily significant unit is a population that is reproductively isolated from other populations of the same species, which therefore represents an important part of the evolutionary history and future evolutionary potential of the species. For example, the

6

species of Pacific salmon are subdivided into many distinct spawning runs that are evolutionarily significant units of central importance for the future survival and evolution of the species (Waples 1991).

New species often arise when genes from two species combine and the number of
chromosomes is increased, a process called polyploidy. Polyploidy has given rise to
many species of plants and some animals, including trout and salmon. Hybrid
populations may play unique ecological roles and may stimulate evolutionary
processes. For example, hybrid populations of plants sometimes provide opportunities
for increased speciation among herbivorous insects (Bush 1975). The biological
processes that produce these genetic mixtures are natural components in the larger
processes of speciation and evolution. For these reasons, it is scientifically
appropriate to protect species of hybrid origin.

2. Science and Listing Priorities.

Currently more than 3,000 species are "Candidates" for listing under the Endangered
Species Act, including more than 2,000 vascular plants, 200 mammals, and 750
insects. This large number of candidate species greatly exceeds the capacity of the
Fish and Wildlife Service and National Marine Fisheries Service to evaluate and
propose species for listing as threatened or endangered. In recent years, about 100
species have been listed annually.

The scarcity of resources available for listing species requires agencies to make
choices about how those resources can best be allocated to meet the objectives of the
Endangered Species Act. In the 1970s and 1980s, the FWS developed several
different schemes for setting priorities for listing species. These priority systems
incorporated such criteria as: magnitude and imminence of threat, availability of
information, taxonomic distinctness of the species, recovery potential, and population
status. The current scheme, adopted in 1983, establishes priorities for listing based
on three criteria: (1) Magnitude of threat, (2) immediacy of threat, and (3) taxonomic
status (the greater the evolutionary distinctness of a taxon, the higher its priority). A
fourth criterion--recovery potential--is included in setting priorities for the development
of recovery plans.

This system of priority-setting has the advantage of being relatively simple. It uses
information that is available for most species, and employs criteria that can be
evaluated relatively objectively (Tobin 1990). However, it does not take full advantage
of ecological knowledge that could better guide limited resources. From an ecological
perspective, three attributes should be considered in a determination of listing

7

(a) Inclusive benefits. Will the habitat managed on its behalf benefit other species, especially species that are listed or are candidates for listing?

Given the limited resources available for endangered species protection, giving high priority to species that serve as protective "umbrellas" for other species makes good ecological sense. For example, the Florida Scrub Jay (Aphelocoma coerulescens coerulescens) is restricted to scrub oak habitats on the Florida peninsula. Many rare species of reptiles, insects, and plants inhabit, and are restricted to, those scrub habitats. Many of them benefit from the land that is managed for the protection of the jay. Similarly, many but not all species requiring old-growth temperate rain forest will benefit if sufficient spotted owl habitat is protected.

The umbrella species approach must be used carefully because every acre of land or boc, of water will contain large numbers of species. Thus, virtually any organism could be considered an umbrella species at some scale. Moreover, an important fact about endangered species is that they rarely have exactly the same requirements. Therefore, even when a suitable umbrella species exists, the ecological needs of other community members must also be considered. The most useful umbrella species are ones whose habitats harbor numerous endemic, rare species. Thus, umbrella species should be given priority for listing in proportion to the number of other endemic, rare species that co-occur with them.

(b) Ecological role. Does the species play an especially important role in the ecosystem in which it lives? Do other species depend on it for their survival? Will its loss substantially alter the functioning of the ecosystem?

Keystone species--an organism whose impact on its community or ecosystem is large, and disproportionately large relative to its abundance (Power and Mills 1995)--ment special attention in the listing process. Unfortunately, determining which species are keystone and which are not is difficult because a species' importance in an ecosystem is not necessarily proportional to its size, abundance, or charisma. Tiny fig wasps and African elephants are both keystone species.

(c) Taxonomic distinctness. How evolutionarily distinct is the taxon in question?

On scientific grounds, the more evolutionarily distinct an organism is, the higher should be its priority for protection. All things being equal, therefore, saving the sole surviving member of a genus may have a higher priority than saving an imperiled species within a large genus that contains many other species. Similarly, protecting full species would normally be given a higher priority than protecting subspecies and populations (Vane-Wright, Humphries, and Williams 1991).

6

Species also have important scientific, aesthetic, and social values, but, given the
paucity of information about most species, priorities are difficult to assign using those
values. Therefore, provisionally it seems scientifically reasonable to give high priority
to species immediately threatened with extinction, to umbrella species, and to
taxonomically unique species. Existing priorities for listing also could be modified by
including considerations of inclusive benefits and ecological role. For example, among
current high priority species (species and monotypic genera facing high magnitude
imminent threats), those providing more inclusive benefits or playing more important
ecological roles should be given higher priority.

B. The Use of Science to Establish Recovery Priorities

The immediate consequence of listing a species under the Endangered Species Act is
to trigger a series of processes that can recover the species and enable it to be
delisted. Recovery is much more complex and difficult than listing, and development
of recovery plans usually requires the generation of substantial new information in
addition to the evaluation of existing information.

1. Science and Critical Habitat Designation.

Once a species is listed, the Endangered Species Act requires the designation of
"critical habitat." Because loss of habitat is the cause of endangerment of most
species, designation and preservation of habitat is a vital part of Endangered Species
Act procedures. Because recovery is a long-term, not a short-term process, and the
goal of the Act is to preserve species in perpetuity, enough habitat must be preserved
to allow the species to survive in the long term. But how long is long term and how
much is enough?

The scientific procedure used to estimate the probability of survival of a population for
a specified period of time is known as Population Viability Analysis, or PVA (Shaffer
1990). Although there is no strict definition of what is or is not included, each PVA
should include an analysis of the best available information on the focal species.
Most PVA analyses combine data from field studies with simulation modeling of the
possible impacts of various extinction factors (Doak et al. 1994, Murphy et al. 1990;
Menges 1990; Stacey and Taper 1992).

The details of a PVA analysis depend on the characteristics of the focal species
(Murphy et al.1990). Species with low population densities and small geographic
ranges (most endangered large vertebrates, for example) and small geographic
ranges (many plants) require a PVA that includes analysis of the genetic and
demographic factors that affect small populations. Smaller organisms, such as most