1: \begin{abstract}

2: The magnifications of the images in a strong gravitational lens

3: system are sensitive to small mass clumps in the lens potential;

4: this effect has been used to infer the amount of substructure in

5: galaxy dark matter halos.  I study the theory of substructure

6: lensing to identify important general features, and to compute

7: analytic cross sections that will facilitate further theoretical

8: studies.  I show that the problem of a clump anywhere along the

9: line of sight to a lens can be mapped onto an equivalent problem

10: of a clump in a simple convergence and shear field; clumps at

11: arbitrary redshifts are therefore not hard to handle in

12: calculations.  For clumps modeled as singular isothermal spheres

13: (SIS), I derive simple analytic estimates of the cross section for

14: magnification perturbations of a given strength. The results yield

15: two interesting conceptual points. First, lensed images with positive

16: parity are always made brighter by SIS clumps; images with negative

17: parity can be brightened but are much more likely to be dimmed.

18: Second, the clumps need not lie within the lens galaxy; they can be

19: moved in redshift by several tenths and still have a significant

20: lensing effect. Isolated small halos are expected to be common in

21: hierarchical structure formation models, but it is not yet known

22: whether they are abundant enough compared with clumps inside lens

23: galaxies to affect the interpretation of substructure lensing.

24: \end{abstract}

25: