obtained from the current HIV literature.

Sousa et al. then built computer simulations to test if an 'ill-adapted SIV' (meaning a simian immunodeficiency virus already infecting a human but incapable of transmission beyond the short acute infection period) could spread in colonial cities. The simulations used parameters of sexual transmission obtained from the current HIV literature. They modelled people's 'sexual links', with different levels of sexual partner change among different categories of people (prostitutes, single women with several partners a year, married women, and men), according to data obtained from modern studies of sexual promiscuity in African cities. The simulations let the parameters (city size, proportion of people married, GUD frequency, male circumcision frequency, and transmission parameters) vary, and explored several scenarios. Each scenario was run 1,000 times, to test the probability of SIV generating long chains of sexual transmission. The authors postulated that such long chains of sexual transmission were necessary for the SIV strain to adapt better to humans, becoming a HIV capable of further epidemic emergence.

The main result was that genital ulcer frequency was by far the most decisive factor. For the GUD levels prevailing in Kinshasa, in early 20^th century, long chains of SIV transmission had a high probability. For the lower GUD levels existing in the same city in the late 1950s (see above), they were much less likely. And without GUD (a situation typical of villages in forested equatorial Africa before colonialism) SIV could not spread at all. City size was not an important factor. The authors propose that these findings explain the temporal patterns of HIV emergence: no HIV emerging in tens of thousands of years of human slaughtering of apes and monkeys, several HIV groups emerging in the nascent, GUD-riddled, colonial cities, and no epidemically successful HIV group emerging in mid-20^th century, when GUD was more controlled, and cities were much bigger.

Male circumcision had little to moderate effect in their simulations, but given the geographical correlation found, the authors propose that it could have had an indirect role, either by increasing genital ulcer disease itself (it is known that syphilis, chancroid, and several other GUDs have higher incidences in uncircumcised men), or by permitting further spread of the HIV strain, after the first chains of sexual transmission permitted adaptation to the human organism.

One of the main advantages of this theory is stressed by the authors: "It [the theory] also offers a conceptual simplicity because it proposes as causal factors for SIV adaptation to humans and initial spread the very same factors that most promote the continued spread of HIV nowadays: promiscuous sex, particularly involving sex workers, GUD, and possibly lack of circumcision