Successful testing depends upon projecting from a test situation to the per
formance of a technology under working conditions. Some have argued that pr
ojection is made possible primarily through a collective agreement that the
test circumstances and working conditions are similar in crucial ways. Thi
s paper argues that projection is better understood in many cases as a set
of local practices through which testing produces change in a wider context
of technology use. In the earthquake-engineering laboratory described here
, this connection depends upon the circulation of skilled people, material
objects and symbolic representations across distinct work settings, in such
a way as to construct a continuous 'chain of practices' between the labora
tory and the worlds of construction, academia and design engineering. This
analysis highlights the importance of the division of labour in shaping tec
hnical practice both inside and outside the laboratory.