In a dihybrid cross,when would the proportion of parental gene combinations be much higher than non-parental types,as experimentally shown by Morgan and his group?

When the genes are linked the proportion of parental gene combinations be much higher than non-parental types.