Mutation accumulation (MA) experiments employ the strategy of minimizing the population size of evolving lineages to greatly reduce effects of selection on newly arising mutations. Thus, most mutations fix within MA lines independently of their fitness effects. This approach, more recently combined with genome sequencing, has detailed the rates, spectra, and biases of different mutational processes. However, a quantitative understanding of the fitness effects of mutations virtually unseen by selection has remained an untapped opportunity. Here, we analyzed the fitness of 43 sequenced MA lines of the multi-chromosome bacterium Burkholderia cenocepacia that had each undergone 5554 generations of MA and accumulated an average of 6.73 spontaneous mutations. Most lineages exhibited either neutral or deleterious fitness in three different environments in comparison with their common ancestor. The only mutational class that was significantly overrepresented in lineages with reduced fitness was the loss of the plasmid, though nonsense mutations, missense mutations, and coding insertion-deletions were also overrepresented in MA lineages whose fitness had significantly declined. Although the overall distribution of fitness effects was similar between the three environments, the magnitude and even the sign of the fitness of a number of lineages changed with the environment, demonstrating that the fitness of some genotypes was environmentally dependent. These results present an unprecedented picture of the fitness effects of spontaneous mutations in a bacterium with multiple chromosomes and provide greater quantitative support for the theory that the vast majority of spontaneous mutations are neutral or deleterious.KEYWORDS fitness effects; deleterious mutation; pleiotropy; genetic drift; Burkholderia cenocepacia T HE extent to which spontaneous mutations contribute to evolutionary change largely depends on their rates and fitness effects. Both parameters are fundamental to several evolutionary problems, including the maintenance of genetic variation (Charlesworth et al. 1993(Charlesworth et al. , 2009Charlesworth and Charlesworth 1998), the evolution of recombination (Muller 1964;Kondrashov 1988;Otto and Lenormand 2002;Roze and Blanckaert 2014), the evolution of mutator alleles (Sniegowski et al. 1997;Tenaillon et al. 1999), and deleterious mutation accumulation (MA) in small populations (Lande 1994;Lynch et al. 1995Lynch et al. , 1999Schwander and Crespi 2009). Many studies have now obtained direct and robust estimates of mutation rates and spectra across diverse organisms, but our understanding of the fitness effects of spontaneous mutations remains limited to mostly indirect estimates in classic model organisms (Eyre-Walker and Keightley 2007).MA experiments provide the opportunity to quantify properties of the fitness of spontaneous mutations that have not been exposed to the sieve of natural selection. Specifically, MA experiments limit the efficiency of natural selection by passaging replicate lineages throu...