Structural Biochemistry/Nucleic Acid/DNA/Meselson-Stahl Experiment
Meselson – Stahl Experiment 
Theories of Replication of DNA 
The daughter DNA is composed entirely of new DNA and the parent DNA retains it’s same back-bones and bases.
Replication produces two copies of DNA that are made up of 50% DNA from the parent DNA helix, and 50% of new DNA. In this situation, each daughter DNA double-helix contains one strand that is the old DNA (from the parent) and one strand that is new (the complimentary strand resulting from the replication).
This form of replication also produces daughter DNA that is constituted by 50% new DNA, and 50% parent DNA. However, in this case, the new DNA and old DNA are shuffled, and fragments of each are found on both strands on the helices on both copies of DNA following replication.
Schematic of the three theories of replication, by CJHIGGIN
The Experiment 
Watson and Crick proposed that DNA replicated semi-conservatively, but conservative and dispersive replication were still plausible until the theories could be disproved. In 1957, Matthew Meselson, and Franklin Stahl devised an experiment to determine whether DNA replicated following a conservative, semi-conservative, or dispersive model.
Meselson and Stahl cultured Escherichia Coli in a medium containing a heavy isotope of nitrogen (15N) as the only nitrogen source, as opposed to the more common nitrogen-14 (14N). After several generations, the E. coli contained DNA composed of nucleotide base made of 15N isotope. The (15N) DNA was denser than the common (14N) DNA, and the difference in densities allowed for separation by density gradient equilibrium sedimentation.
To achieve separation of the E. coli DNA by densities, the DNA was mixed with a solution of CsCl and centrifuged. A CsCl density gradient was created as a result of sedimentation and diffusion working against each other. The DNA molecules were found in the area of the CsCl density gradient that was equal to their own density.
The (15N) E. colicells were transferred to a medium that contained only (14N). DNA was isolated from the first generation of cells grown in the (14N) medium, and analyzed by density gradient equilibrium sedimentation. Then DNA from the second generation of E. coli grown in the (14N) medium was extracted and analyzed.
The first generation of E. coli grown in the 14N medium contained a single DNA band found halfway in between where the 14N DNA band and the 15N DNA band should have been. This demonstrated the presence of a DNA that was lighter than the DNA from the original population of E. Coli grown in the 15N medium, but still heavier than 14N DNA. Due to the position of this intermediate DNA band in the density gradient, it was apparent that the DNA was a hybrid and contained both 14N and 15N. This automatically eliminated the conservative model of replication, which would have resulted in two distinct bands: one matching the position of the 15N-containing DNA, and one matching the position expected by DNA containing only 14N. Only the dispersive and semi-conservative models fit the situation.
The second generation of E. coli grown in the 14N medium contained two distinct bands. One of the bands was 14N DNA, and the other band was the intermediate (14N/15N) DNA. This result supported the theory of semiconservative replication since dispersive replication would have resulted in a single band of lower density DNA in each consecutive generation.
The figure below illustrates the theoretical outcome of the conservative, dispersive, and semiconservative models along with the experimental outcome obtained by Meselson and Stahl.
Figure: A schematic of the appearance of fractions of DNA samples after centrifuging in a density gradient, by CJHIGGIN
Berg, Jeremy M., John L. Tymoczko, and Lubert Stryer. "Exploring Genes and Genomes." Biochemistry. New York: W. H. Freeman, 2007. 113-14. Print.
Campbell and Reese's Biology, 7th Edition
Nelson and Cox's Lehninger Principles of Biochemistry, 5th Edition