Fig. 3.—

Preferential loss of small introns. A boxplot of the sizes of introns that were either 100% retained or found to be lost in at least one (A) Caenorhabditis or (B) Drosophila species. For the lost introns, we plotted the size of the introns that were retained at identical positions in neighboring species, excluding initial introns that possibly contain indispensable regulatory elements in the often larger introns. The median of introns that are lost was significantly smaller than that of retained introns for all Caenorhabditis (***P < 0.001) and Drosophila species (***P < 0.001, Mann–Whitney U test). For C. elegans: n = 97,220 for retained introns; n = 10,465 for lost intron. For Drosophila: n = 142,967 for retained introns; n = 3,274 lost introns.