I take it that the author was not quite right there, and that mutations tend not to be as commonly harmful as he suggests. Otherwise we would not see organisms evolving as they have done.
He was surprisingly accurate. Most genetic mutations that have an effect on the phenotype...on the biochemistry, are harmful. A suitably complex biochemical structure is much more likely to be ruined by even a slight change than it is to be improved. Populations tend to find themselves either evolving in light of environmental change, or more generally they find themselves in equilibrium - all the phenotypes are generally clustered around an optimum, and deviating from the optimum is likely to result in selection culling things back. It's only when the optimum changes that we see the mutations that might make the organism less badly adapted for the new environment.