This paper considers estimation based on a set of T + 1 discrete obser
vations, y(0),y(h),y(2h),...,y(Th) = y(N), where h is the sampling fre
quency and N is the span of the data. In contrast to the standard appr
oach of driving N to infinity for a fixed sampling frequency, the curr
ent paper follows Phillips [35,36] and Perron [29] and examines the "d
ual" asymptotics implied by letting h tend to zero while the span N re
mains fixed. We suggest a way of explicitly embedding discrete process
es into continuous-time processes, and using this approach we generali
ze the results of the above-mentioned authors and derive continuous re
cord asymptotics for vector first-order processes with positive roots
in a neighborhood of one and we also consider the case of a scalar sec
ond-order process. We illustrate the method by two examples. The first
example is a near unit root model with drift and trend. We derive the
continuous record approximation to the Dickey-Fuller rho(tau) triple-
overdot test and to a recent test by Schmidt and Phillips [39] and tab
ulate the distribution of the test statistics in the near continuous r
ecord situation. The second example is a (near) I(2)-process. We prese
nt a continuous record approximation to the least-squares estimator an
d compare to recent results of Perron [32].