Population bottlenecks can have major effects in the evolution of RNA viruses, but their possible influence in the evolution of DNA viruses is largely unknown. Genetic and biological variation of herpes simplex virus type 1 (HSV-1) has been studied by subjecting 23 biological clones of the virus to 10 plaque-to-plaque transfers. In contrast to large population passages, plaque transfers led to a decrease in replicative capacity of HSV-1. Two out of a total of 23 clones did not survive to the last transfer in 143 TK -cells. DNA from three genomic regions (DNA polymerase, glycoprotein gD and thymidine kinase) from the initial and passaged clones was sequenced. Nucleotide substitutions were detected in the TK and gD genes, but not in the DNA polymerase gene. Assuming a uniform distribution of mutations along the genome, the average rate of fixation of mutations was about five mutations per viral genome and plaque transfer. This value is comparable to the range of values calculated for RNA viruses. Four plaque-transferred populations lost neurovirulence for mice, as compared with the corresponding initial clones. LD 50 values obtained with the populations subjected to serial bottlenecks were 4-to 67-fold higher than for their parental clones. These results equate HSV-1 with RNA viruses regarding fitness decrease as a result of plaque-to-plaque transfers, and show that population bottlenecks can modify the pathogenic potential of HSV-1. Implications for the evolution of complex DNA viruses are discussed. INTRODUCTIONThe concept that small asexual populations of organisms will tend to accumulate deleterious mutations unless genetic lesions are repaired by sex or recombination was first proposed by Muller (Muller, 1932, 1964 and is known as the Muller's ratchet hypothesis (Bell, 1988;Colato & Fontanari, 2001;Felsenstein, 1974;Maynard Smith, 1976;Nowak & Schuster, 1989). Its operation has found experimental confirmation mainly in studies carried out with RNA viruses (Chao, 1990; Clarke et al., 1993;de la Iglesia & Elena, 2007;Duarte et al., 1992Duarte et al., , 1994Escarmís et al., 1996Escarmís et al., , 2009Novella, 2004; Novella & EbendickCorpus, 2004; reviewed by Escarmís et al., 2006), and also with plants, protozoa, mitochondrial DNA and bacteria (Allen et al., 2009;Andersson & Hughes, 1996;Bell, 1988;Coates, 1992; Engelstädter, 2008;Loewe, 2006;Moran, 1996). The effects of Muller's ratchet are accentuated when an asexual genome population undergoes severe bottleneck events. In the case of RNA viruses, mutation rates in the range of 10 23 to 10 25 substitutions per nucleotide copied (Batschelet et al., 1976;Domingo et al., 1996Domingo et al., , 2006Drake & Holland, 1999; Sanjuán et al., 2010) dictate that 0.1 to 10 mutations will occur upon synthesis of any new genomic molecule in an infected cell. Generally, the frequency of deleterious mutations is on average higher than the frequency of beneficial mutations. In consequence, the random sampling of a genome from a population to give rise to a new plaque on a cell...