the city: good for migrants, bad for genes?

Arnold Kling and Richard Florida have recently started up blog conversation on whether cities are creative growth centers or urban death traps.

Kling quotes Razib Khan:

up until the year 1900 the world's cities were massive genetic blackholes. Cities only kept their population up through migration, which explains how Rome shrunk to 30,000 inhabitants by the 7th century.

Kling asks how this jives with Jane Jacobs, who contends that urban settlements have been the catalysts of economic growth. Jacobs isn't directly contradicted by the genetic data, but the two certainly appear to point in different directions. If cities were genetic graveyards up until 1900, how can it be that cities were simultaneously hubs of economic growth?

One way to reconcile the two perspectives is to restate Jacobs' position: when circumstances allow for economic growth, returns on agglomeration will concentrate this growth in urban settlements, which will in turn spur further development. Population density is not a guarantor of growth. Furthermore, when lacking the conditions necessary for growth, urban settlements' negatives, such as disease, may outweigh any benefits to agglomeration, leading to a relatively poor genetic expectancy.

The Jacobs question, however, really only touches on a more fundamental question: if cities were genetic graveyards, why did migrants continue to flock to these urban death traps?

Why were genetic returns to agglomeration negative during a period when economic returns were relatively high, according to Jacobs and historical migration patterns? What in the name of Charles Darwin was going on?

I'm interested in hearing theories (perhaps the immediate prospects for the migrant is better in the city, but their genes are more likely to be wiped out generations later by a plague...), but for now, I am skeptical that millions of people, throughout history, have lowered their genetic expectancy by migrating to urban death traps.