1: \begin{abstract}
2: We explore the origin of flux ratio anomaly in quadruple lens systems.
3: Using a semi-analytic method based on
4: $N$-body simulations, we estimate the effect of possible magnification
5: perturbation caused by subhaloes with a mass scale of $ \lesssim 10^9\,h^{-1} \ms$ in lensing
6: galaxy haloes. Taking into account astrometric shifts, assuming that the primary lens is
7: described by a singular isothermal ellipsoid, the expected change to the flux ratios per a multiply
8: lensed image is just a few percent and the mean of the expected convergence
9: perturbation at the effective Einstein radius of the
10: lensing galaxy halo is $\langle \delta \kappa_{\tr{sub}} \rangle = 0.003$, corresponding to
11: the mean of the ratio of a projected
12: dark matter mass fraction in subhaloes
13: at the effective Einstein radius $\langle f_{\tr{sub}} \rangle = 0.006$.
14: In contrast, the expected change to the flux ratio
15: caused by line-of-sight structures is typically $\sim 10$ percent and
16: the mean of the convergence perturbation is
17: $\langle |\delta \kappa_{\tr{los}}| \rangle = 0.008$, corresponding to $\langle
18: f_{\tr{los}} \rangle = 0.017$.
19: The contribution of magnification perturbation caused
20: by subhaloes is $\sim 40$ percent of the total at a source redshift
21: $z_\tr{S}= 0.7$ and decreases monotonically in $z_\tr{S}$ to $\sim
22: 20$ percent at $z_\tr{S}= 3.6$. Assuming statistical isotropy, the convergence
23: perturbation estimated from observed 11 quadruple lens systems has
24: a positive correlation with the source redshift $z_\tr{S}$, which is much stronger
25: than that with the lens redshift $z_{\tr{L}}$. This feature also supports an idea
26: that the flux ratio anomaly is caused mainly by
27: line-of-sight structures rather than subhaloes. We also discuss about
28: a possible imprint of line-of-sight structures in demagnification
29: of minimum images due to locally underdense structures in the line of sight.
30: \end{abstract}
31: