The last number of The Economist has an article entitled "Bogus Academia" about the rise of predatory journals. The article informs that it took 300 years for science to establish 10,000 reputable journals, while predatory journals managed to amass the same quantity in just 20 years.
The article shows that the problem is not the same for different scientific fields. While in the field of Economics predatory journals represent just 30% of all journals. In the field of Health predatory journals represent 43%. However, 61% of journals in the field of Math, physical and chemical sciences are of the predatory type!
According to the article, a notorious difference between reliable and predatory journals is the percentage of papers that never get cited. 10% on the former and 60% on the latter.