1: \begin{abstract}

2: The method of Helsing and co-workers evaluates Laplace and related layer potentials generated by a panel (composite) quadrature on a curve,

3: efficiently and with high-order accuracy for arbitrarily close targets.

4: %

5: Since it exploits complex analysis, its use has been restricted to two dimensions (2D).

6: %

7: We first explain its loss of accuracy as panels become curved,

8: using a classical complex approximation result of Walsh

9: that can be interpreted as ``electrostatic shielding''

10: of a

11: %the panel's

12: Schwarz singularity.

13: %

14: We then introduce a variant that swaps

15: the target singularity for one at its

16: {\em complexified parameter preimage};

17: in the latter space the panel is flat, hence the

18: convergence rate can be much higher.

19: %increasing efficiency.

20: % accuracy?

21: %

22: The preimage is found robustly by %a simple

23: Newton iteration.

24: %

25: This idea also enables, for the first time, a near-singular quadrature

26: for potentials generated

27: by smooth curves in 3D, building on recurrences of Tornberg--Gustavsson.

28: %

29: We apply this to accurate evaluation of the Stokes flow near to a

30: %loop-like

31: curved filament in the slender body approximation.

32: %

33: Our 3D method is several times more efficient (both in terms of

34: kernel evaluations, and in speed in a C implementation) than

35: the only existing alternative, namely, adaptive integration.

36: \end{abstract}