Enhancers could, in principle, function by increasing the rate of reinitiation on individual adjacent active promoters or by increasing the probability that an adjacent promoter is activated for transcription. We have addressed this issue for the repetitive metazoan rDNA enhancer by microinjecting Xenopus oocytes with enhancer-less and enhancer-bearing genes and determining by EM the frequency that each gene type forms active transcription units and their transcript density. We use conditions where transcription requires the normal rDNA promoter and is stimulated 30-50-fold by the enhancer. (In contrast, at saturating template conditions as used in previous EM studies, an aberrant mode of transcription is activated that is not affected by the rDNA enhancer or by the generally recognized rDNA promoter). The active transcription units on enhancer-less genes are found to be as densely packed with nascent transcripts and polymerases as those on enhancer-bearing genes and on the endogenous rRNA genes. Significantly, the enhancer-bearing genes are approximately 30-50-fold more likely to form such active transcription units than enhancer-less genes, consistent with their amounts of transcript. Complementary studies confirm that the enhancer does not affect elongation rate, the stability of the transcription complex, or transcript half-life. These data demonstrate that the repetitive metazoan rDNA enhancer causes more genes to be actively transcribed and does not alter the reinitiation rate on individual active genes.