Counting scats (excrement) of wildlife species like deer and carnivores can provide an index to relative density of the species but there are too many unknown factors and variables to obtain a solid estimate. Thus you could count scats in one area and compare that number to the number of scats in another area (with the same environmental conditions) and you could reasonably conclude that the density of a species in the first area was different than the density in another area. But it would very difficult to convert these numbers to density, that is number of foxes per square kilometre.
Some of the problems with using scats to estimate true density is:
1. scats deteriorate at different rates depending on weather. Heavy rain will wash the fine materials from scats of carniovres like foxes and then only bone chips, hair and berry pits are left.
2. the number of scats encountered could differ simply because of terrain features even though the density of foxes were the same. For example in uneven terrain foxes would likely confine their travel routes to a few trails but on flat ground they may have many more routes available. Hence, in the uneven terrain environment the number of scats along the trails would be greater than that along the many trails in the even terrain environment. If you used number of scats seen per 1000 meters of trail walked then you would over estimate the density of foxes in the uneven terrain and under estimate the density of foxes in the even terrain.
So for general observation that there are more foxes in one place than another, scat counts would be okay to use as a broad estimate of relative abundance. However, for true density a fairly complicated study design would have to be developed involving a rigorous sampling scheme, measurement of other indices of abundance of foxes, and accounting for the various biases.