On this page
Every NEET aspirant learns that DNA replication is semiconservative. Far fewer can explain how anyone knew. The answer is one experiment, done at Caltech by two young scientists, that turned a guess from the Watson and Crick model into a measured fact. This post tells you who they were, what they actually did, and how to solve the generation-counting questions that come straight out of it.
Who were Meselson and Stahl?
According to Philip Hanawalt's historical commentary in PNAS (2004), Matthew Meselson began graduate study in chemistry at Caltech in 1953 in the laboratory of Linus Pauling. He spent the summer of 1954 at the Marine Biological Laboratory at Woods Hole, where he met Frank Stahl, a biology graduate student from the University of Rochester. Meselson helped Stahl with a calculus problem in phage genetics, and the two began talking about how to test DNA replication. Stahl was already heading to Caltech as a postdoc, so the collaboration continued there.
Their key tool came from Meselson's thesis work with Jerry Vinograd. Hanawalt describes how, when DNA was spun at high speed in a concentrated caesium chloride (CsCl) solution, the salt itself formed a density gradient and the DNA gathered into a narrow band at the point where its density matched the salt. Heavier DNA would band lower; lighter DNA, higher.
What did the experiment actually do?
The procedure, as summarised in the original paper, "The replication of DNA in Escherichia coli" (PNAS, 1958) and in Hanawalt's account, had three steps:
Grow E. coli for many generations in a medium where the only nitrogen source is 15NH4Cl, so its DNA becomes fully "heavy".
Switch the bacteria to ordinary 14NH4Cl and keep them growing.
Take samples as the cells divide, extract the DNA and spin each sample in a CsCl gradient to see where it bands.
The NCERT Class 12 Biology chapter on the molecular basis of inheritance reminds you that 15N is not radioactive. It is told apart from 14N only by density.
The results were clean. At one generation, Hanawalt writes, "only the intermediate density hybrid band was present." After a second generation there were equal amounts of hybrid and light DNA. The original paper put the conclusion carefully: the nitrogen of a DNA molecule is divided equally between two subunits that remain intact through many generations.
How does the result rule out the other models?
There were three serious possibilities. Write down what each predicts, and the experiment does the rest.
Conservative: the parent double helix stays intact and a wholly new helix is made. After one generation you would see two bands, heavy and light, and no hybrid. The single hybrid band rules this out.
Dispersive: old and new DNA are mixed in patches along both strands. After one generation you would see one intermediate band, which matches. But after two generations you would still see one band, just a little lighter, not a separate light band. The appearance of a distinct light band rules this out.
Semiconservative: each daughter helix has one old strand and one new strand. One hybrid band after one generation; hybrid plus light in equal amounts after two. This is what they saw.
Meselson and Stahl also heated the DNA to separate its strands. Hanawalt reports that after holding samples at 100 °C for 30 minutes in CsCl, the hybrid DNA split into two bands at the positions of heat-denatured 15N DNA and 14N DNA. Each hybrid molecule really was one heavy part plus one light part.
Worked example: the NCERT 80-minute question
Your NCERT textbook says E. coli divides every 20 minutes in this setting and asks: if the culture grows for 80 minutes in 14N, what are the proportions of light and hybrid DNA?
Start from one heavy DNA molecule and follow it.
80 minutes divided by 20 minutes per generation gives 4 generations.
After n generations one molecule has become 2^n molecules. Here 2^4 = 16 molecules.
The two original heavy strands never break, so exactly 2 of those 16 molecules carry an old strand. Those 2 are hybrid.
The rest, 16 − 2 = 14, are light.
So hybrid : light = 2 : 14 = 1 : 7. Hybrid DNA is 2/16 = 12.5% of the total. You can check the rule against the real data: at 2 generations it gives 2 hybrid out of 4, the 1:1 mix Meselson and Stahl reported.
Common mistakes to avoid:
Counting strands instead of double-stranded molecules. The question asks about DNA molecules.
Using 80 minutes as 8 generations. Divide by the doubling time first.
Forgetting that the number of hybrid molecules stays at 2 forever, while light molecules keep increasing. Hanawalt notes this "forever" feature too.
What can a student take from Meselson and Stahl's way of working?
Design the result before you run it. They could only read their bands because they knew what each model predicted. Before any experiment question, list the predictions of every model first.
Do the step that could prove you wrong. Hanawalt stresses that continuing to a second generation was essential. One clean result is not enough if another model also explains it.
State only what you have shown. Their paper spoke of "subunits", not single strands, because their data did not yet prove that. Precise wording is a scientific skill, and it earns marks.
If the difference between conservative, semiconservative and dispersive copying still feels shaky, the Fermi Concept Library covers every concept in physics, chemistry, biology, mathematics and computing, so you can revisit one idea at a time. For the enzymes that carry out the copying, read our guide to DNA replication and transcription and the wider chapter on the molecular basis of inheritance.
Next step: draw the band pattern for generations 0 to 4 under all three models from memory, then check it against this post.
