Virology was born in 1898, and has suffered from sampling bias ever since. For decades, viruses were defined by what they were not: not as big as a bacterium, not visible with a microscope, not culturable in the absence of a host. At the dawn of the 20th century, undiscovered viruses and aspiring Hollywood starlets had the same challenge—how to catch someone’s attention. Thus, the first viruses known to science were those causing very obvious phenotypes in clinically or agriculturally important hosts: yellow fever in humans, mosaic disease in tobacco, foot-and-mouth disease in livestock. The acid-test was the ability to pass through the pores of a Chamberland or Berkefeld filter; the viruses that were easiest to find were those that produce infectious cell-free virions—another source of discovery bias. And, in the days before cell culture, the easiest of those to study were the ones that induced rapid and obvious pathogenesis in a conveniently available host organism. By mid-20th century, technology fostered the move to reductionist experimentation, including electron microscopy, X-ray crystallography, cell-culture and, of course, recombinant DNA technology. Virologists moved into laboratories, and the model systems they chose to bring with them were often those already in place, proven to be tractable and considered relevant by funding agencies. Today you can peruse any virology textbook and get the impression that less than two dozen viruses represent the entirety of Earth’s virosphere.
Click "source" to read more.