A new method is presented and used to determine the IMF of the starburst cl
uster NGC2070. A new correction, the magnitude-limit correction is introduc
ed, and shown to be crucial when attempting to derive the IMF in the presen
ce of variable reddening when the photometry is not several magnitudes deep
er than the fainter stars analyzed. Failure to apply this correction is res
ponsible for the drop at the low mass end of the IMF found in previous work
on this cluster, despite the proper application of incompleteness correcti
ons. For masses between 3M(.) less than or similar to M less than or simila
r to 120M(.) and outside 15" the IMF of NGC2070 is shown to be consistent w
ith being a single power law with a Salpeter exponent. In the central regio
n (4.6" < r < 19.2") within 2.8M(.) < M < 120M(.) our data combined with HS
T observations yield a slope flatter than Salpeter at the 2-3 sigma level.
Furthermore, it is shown that the number of M > 50M(.) stars near the core
(Massey & Hunter 1998a, 1998b) is incompatible with the intermediate mass c
ounts of Hunter et al. (1995, 1996) extrapolated with a Salpeter slope, so
either the slope is flatter than Salpeter, or the HST spectral types are bi
ased towards earlier types. The star-formation history is dominated by thre
e bursts of increasing strength occurring 5My, 2.5My, and less than or simi
lar to 1.5My ago, the latest one responsible for most of the star-formation
within 6 pc from the cluster center. A spherically symmetric structure is
detected at about 6 pc from the cluster center which contains predominantly
massive stars and has a flatter IMF The surface number density profile of
the cluster is shown to be well modeled by a single power law, Sigma(R) sim
ilar or equal to R-alpha, over 0.4 pc < R < 12 pc, with alpha approximate t
o 1.85, significantly steeper than isothermal.