106 / Chapter 4
document that existing animal studies had been fully evaluated for
validity and clinical generalizability would reduce the numbers of
animals used.
These two steps aim to reduce the numbers of animals in labs
and ensure that, even in the event of a disaster, all animals affected
will have at least the minimal federal protection. But a third step is
needed: research facilities must take the lessons of recent disasters
seriously and address worst-case scenarios in their planning. It is
not enough to move the animals to upper fl oors if the emergency
generators are in the basement. It is not enough to have a plan to
evacuate animals in case of fl ood, if the facility sits in a fl ood zone
that may become inaccessible to the people responsible for staging
the evacuation. What is enough will vary by facility, location, and
potential hazards. In the ideal, we would not build research facil-
ities on fl ood plains or earthquake fault lines. In reality, we must
fi nd ways of anticipating and addressing the consequences that
doing so has for the animals we make vulnerable. In the conclud-
ing chapter, I discuss some of the barriers to thinking about worst-
case scenarios and suggest some ways around them.