1: \begin{abstract}
2: The structural evolution of substructure in cold dark matter (CDM)
3: models is investigated combining ``low-resolution'' satellites from cosmological $N$-body
4: simulations of parent halos with $N=10^7$ particles with high-resolution
5: individual subhalos orbiting within a static host potential.
6: We show that, as a result of mass loss, convergence
7: in the central density profiles requires the initial satellites to be resolved
8: with $N=10^7$ particles and parsec-scale force resolution.
9: We find that the density profiles of substructure halos can be well fitted with
10: a power-law central slope that is unmodified by tidal forces even after the tidal
11: stripping of over 99\% of the initial mass and an exponential cutoff in the outer parts.
12: The solution to the missing-satellites problem advocated by Stoehr \etal in 2002
13: relied on the flattening of the dark matter halo central density cusps by gravitational
14: tides, enabling the observed satellites to be embedded within dark halos with maximum
15: circular velocities as large as $60\kms$. In contrast, our results suggest that tidal
16: interactions do not provide the mechanism for associating the dwarf spheroidal
17: satellites (dSphs) of the Milky Way with the most massive substructure halos expected in a
18: CDM universe. Motivated by the structure of our stripped satellites, we
19: compare the predicted velocity dispersion profiles
20: of Fornax and Draco to observations, assuming that they are
21: embedded in CDM halos. We demonstrate that models with isotropic and
22: tangentially anisotropic velocity distributions for the stellar component
23: fit the data only if the surrounding dark matter halos
24: have maximum circular velocities in the range $20-35 \kms$.
25: If the dSphs are embedded within halos this large then the overabundance
26: of satellites within the concordance {\LCDM} cosmological model is
27: significantly alleviated, but this still does not provide the entire solution.
28: \end{abstract}
29: