The real centre of the later city consisted of the _Capitoline Hill_, which, though not of great circumference, is properly composed of two hills, a southern one towards the Forum, and a northern one; between them a considerable depression of the ground is still visible. This depression, however, was far greater in ancient times than it is now, and in it there was a portico open on both sides, but at present its back is filled up with rubbish, especially from the ruins of the Capitoline temple, which, like many other buildings, has been purposely and barbarously destroyed. There was a _clivus_ leading up the Capitoline hill from the Forum, which, as in the case of all the Roman hills, formed an inclined plain ascending gradually. The names of the _clivi_ of all the other hills, however, are not known. On the Quirinal I do not find a _clivus_, but it had a _semita_. The meaning of this latter term is not correctly given in our dictionaries: the _semita_ does not differ so much from a carriage road by being less in breadth, but it is altogether a way which no vehicles can pass, either from its want of breadth, or from its construction in other respects, and which therefore is available only to foot-passengers and mules; _semitae_ were ways like the one still existing in the Vatican palace, by which the pope can ride on a mule into his own apartment. In Germany there is nothing comparable to it; the Italian name is _cordonata_, and it must be conceived as a strongly, though not inconveniently, inclined plain, with high stones at certain intervals for the purpose of stopping, so that the second step begins lower than the point at which the first left off.[17] Semitae are also found at gates, especially of Cyclopean towns, as at Ferentino. Before the time of Trajan there is no trace of a clivus leading up the Quirinal; on the Esquiline I can prove its existence; the Palatine had two clivi, the Aventine one, etc. Rome was essentially different from large modern cities which always contain main streets running from one end to the other; such a street cannot be shown to have existed at Rome, which altogether had but few great streets. All the houses built on the same hill formed, as it were, a small town by themselves with little, and probably extremely irregular, streets, and thus every hill was isolated. It was only the plains and valleys that contained some large streets. The _Esquiliae_ were not a separate street, and the _Carinae_ near the Esquiline also were a quarter of the city rather than a street; the _Subura_ beyond the Esquiliae was a real street, and so also the _Via Sacra_ up to a certain point, but it was not a main street.