1: \begin{abstract}

2: The two-step time discretization proposed by Dahlquist, Liniger and

3: Nevanlinna is variable step $G$-stable. (In contrast, for increasing time

4: steps, the BDF2 method loses A-stability and suffers non-physical energy

5: growth in the approximate solution.) While unexplored, it is thus ideal for

6: time accurate approximation of the Navier-Stokes equations. This report

7: presents an analysis, for variable time-steps, of the method's stability and

8: convergence rates when applied to the NSE. It is proven that the method is

9: variable step,

10: unconditionally, long time stable and second order accurate. Variable step

11: error estimates are also proven. The results are supported by several

12: numerical tests.

13: \end{abstract}

14: