The Poetry of Scientific Experiments

This post is written by UW grad student Sonia Singhal

TL;DR: Like poems, “beautiful” scientific experiments have a cohesive, coherent structure where each part reinforces the whole. In this post, I analyze the structures of the poem “Easter Wings” by George Herbert and the Meselson-Stahl experiment from biology. This work is the opinion of the author, and does not necessarily reflect the views of BEACON or its researchers.

Throughout history, people have said that science removes the beauty from the world. Yet scientists often find great beauty in their work. Scientific experiments have even been called “beautiful” by other scientists. Why is this? What is it about certain experiments that gives the perception of beauty?

I propose that one element may lie in the structure of the experiments. Specifically, we perceive beauty when the individual pieces reinforce each other to form a cohesive, coherent whole. In this way, a “beautiful” scientific experiment is akin to a work of poetry.

In poetry, perhaps more than in any other type of writing, structure (or, as a poet would say, form) is paramount. Poetry is judged not only on what the poet says, but also on how she says it. Aspects of the how include how a reader would recite the poem – for example, the rhyme scheme and the rhythm – as well as the look of the poem on the page. Are the lines short or long? Are they on the left-hand side of the page or the right, or do they straddle the center line? Even white space is meaningful.

A concrete poem, whose shape reflects its subject, takes form to an extreme. I’ll use the concrete poem “Easter Wings,” written by George Herbert in the sixteenth century, to illustrate why structure is so important in poetry.

Lord, who createdst man in wealth and store,
Though foolishly he lost the same,
Decaying more and more,
Till he became
Most poore:
With thee
O let me rise
As larks, harmoniously,
And sing this day thy victories:
Then shall the fall further the flight in me.

My tender age in sorrow did beginne
And still with sicknesses and shame.
Thou didst so punish sinne,
That I became
Most thinne.
With thee
Let me combine,
And feel thy victorie:
For, if I imp my wing on thine,
Affliction shall advance the flight in me.

Herbert’s poem describes the second coming of Christ (a popular subject for writers of the time). However, the beauty of the poem does not come (solely) from its subject. It also comes from the way the poem’s structure reinforces its message and themes.

Since this is a concrete poem, let’s start with the general shape. When you look at the poem from a distance, without reading the words, what does the shape remind you of? (Hint: Tilt your head sideways.)

Herbert’s poem has two stanzas (blocks of text separated by white space), each with a winged, bird-like shape. Herbert achieves this visual effect by altering the length of the lines. He starts the stanza with a line of ten syllables. Each subsequent line has fewer and fewer syllables, until the lines in the middle of the stanza have only two syllables each. Herbert then increases the number of syllables per line until he returns to ten syllables in the final line of the stanza.

When we pair the shape with the text, we find that where the lines contract and expand is not arbitrary. The lines contract when Herbert talks about mankind or himself, made “less” (according to Christian doctrine) by the curse of original sin. In particular, the shortest lines contain words that denote dearth or scarcity: “poore” and “thinne.” One interpretation of the short lines might be that they represent the narrowness or short-mindedness that people can exhibit when they only think about themselves. In contrast, the lines expand when Herbert talks about his faith in God and Christ, which makes him part of something greater than himself.1

Herbert repeats structural elements within his poem. Most obviously, there are two stanzas, each with the appearance of a flying bird. Herbert repeats words and rhymes between the stanzas. Rhymes ending in “-ame,” such as “shame” and “same,” appear in the first half of both stanzas, while rhymes ending in “-ee,” such as “thee” and “me,” appear in the second half.2 The second half of both stanzas also includes “victory,” “flight,” and other words that relate to birds (“larks” in the first verse; “wing” and “imp” in the second. “Imp” is a falconry term – when you imp a bird with broken or clipped wings, you attach new feathers to its wings to allow it to fly again). The references to flight coincide with Herbert’s recovery of hope and faith. Even the last lines of both verses are structurally similar (compare “Then shall the fall further the flight in me” to “Affliction shall advance the flight in me”). The repetition makes the poem feel tightly knitted and could represent the strength of Herbert’s faith.

In “An Essay on Criticism,” eighteenth-century poet Alexander Pope said of poetry that “The sound must seem an echo to the sense.” In other words, structure in poetry must emphasize the meaning. Herbert’s poem exemplifies this philosophy. The bird-like shape of the poem emphasizes the themes of rebirth and renewed faith, while the tight repetition of words and phrases creates a sense of safety and security. The structure and the meaning of the poem resonate with one another, and this resonance makes the poem “beautiful.”

The value of structure extends beyond poetry. The scientific process has its own inherent structure:

  1. You ask a question.
  1. You form a hypothesis, or a reasonable explanation, about the answer.
  1. You run an experiment.
  1. The experiment gives you data (results) that either support or refute your hypothesis.

When each of these pieces (question, hypothesis, experiment, results) reinforces the others with no extraneous content – in other words, when the structure of the whole experiment echoes its core message – the experiment may resonate with us in the same way that a well-constructed poem does.

To explore this idea, I’ll use a classic experiment in biology, the Meselsohn-Stahl experiment, which was fundamental to our understanding of how DNA works.

DNA is a molecule found in all living things. Each time a cell divides, or a parent has a child, the parent’s DNA gets copied and passed on to the next generation. Because the building blocks of DNA were relatively simple compared to other molecules, its importance was originally underestimated. But in the first half of the twentieth century, the results of biologists’ experiments began to indicate that DNA was in fact necessary for life. It encoded information about whether a living being was human, animal, or plant, what it looked like; and how its body worked.

DNA’s shape was first proposed in 1953, by James Watson and Francis Crick.3 At the time, they knew that the building blocks of DNA included four different bases that contained nitrogen (specifically, adenine, cytosine, thymine, and guanine). Watson and Crick also knew that, in any particular DNA molecule, there were equal amounts of adenine and thymine bases, and equal amounts of cytosine and guanine bases, but different amounts of adenine/thymine versus cytosine/guanine. Based on a photograph of X-rays bouncing off DNA taken by Rosalind Franklin, Watson and Crick suggested that DNA was made of two strands that wound around each other in a double helix (Fig. 1). The bases were arranged in pairs between the two strands like rungs on a ladder: adenine with thymine, cytosine with guanine.

Figure 1. Basic shape of a DNA molecule. The two strands (yellow) form a double helix. Between them, adenine (green) pairs with thymine (purple), and cytosine (red) pairs with guanine (blue).4

This arrangement immediately suggested a way of copying DNA without losing the information it encoded. Because the bases are paired, it is possible to determine the sequence of one DNA strand from the other’s: A thymine on one strand always corresponds to an adenine on the other, while a guanine on one strand always corresponds to a cytosine on the other. If you separate the strands, you can make an exact copy of each one to get a new DNA molecule.

However, base pairing also meant that different researchers came up with different ideas on exactly how the process of replication, of copying and creating a new DNA molecule, might work. They suggested three different hypotheses (Fig. 2):

  1. Semiconservative replication. The original DNA strands separated and were copied, and the new DNA molecules were made of one strand of original DNA and one strand of new DNA.
  1. Conservative replication. The original DNA strands separated and were copied, but afterwards the original DNA strands re-paired, and the newly made DNA strands paired with each other. One molecule was made up only of the original DNA, and the other was made up only of new DNA.
  1. Dispersive replication. The process of copying DNA involved making short segments from both strands – in other words, copying some of one strand, then some of the other, then more of the first. Molecules made this way would contain some original DNA and some new DNA in both strands.

Figure 2. Illustration of three hypotheses for how DNA might replicate.5

In 1957, Matthew Meselson, a graduate student, and Frank Stahl, a post-doctoral researcher, designed a simple experiment with bacteria to determine which of these hypotheses was correct.6 They took advantage of the fact that nitrogen, which appears in the DNA bases, has a heavy form and a light form. Normally, the lighter form of nitrogen appears in living things, so using the heavy form of nitrogen in DNA would let Meselson and Stahl track specific strands. In their experiment, they put the heavy form of nitrogen into the DNA of bacterial parents, but they only let the bacterial children and grandchildren use the light form of nitrogen to make DNA copies. The weight of the resulting DNA would reveal how it had been copied.

  1. In semiconservative replication, the weight of the DNA would change with each generation. In the first (child) generation of bacteria, new DNA molecules in the child bacteria would be made of one strand of old (heavy) DNA and one strand of new (light) DNA. DNA in both parents and children would have an intermediate weight.

    In the second (grandchild) generation of bacteria, new DNA molecules could be copied from both the old (heavy) and new (light) strands. There would be a mixture of intermediate-weight DNA (copied from heavy strands) and light-weight DNA (copied from light strands).

  1. In conservative replication, the parent would still have only heavy DNA, and the children and grandchildren would have only light DNA.
  1. In dispersive replication, all new DNA molecules would have some old DNA and some new DNA. All molecules would be of intermediate weight in parents, children, and grandchildren.

When Meselson and Stahl weighed the DNA before and after the bacterial parents reproduced, the results were indisputable. Before reproducing, the parents only had heavy DNA. After the first generation, parents and children had DNA of equal, intermediate weight. After the second generation, Meselson and Stahl saw both intermediate-weight and light-weight DNA. DNA was being copied in a semiconservative manner.

Even before the results of the experiment had been formally published, scientists called Meselson’s and Stahl’s experiment “beautiful.” Meselson said the results were “clean as a whistle.” Maurice Wilkins, a physicist and molecular biologist, described their paper on the experiment as “elegant and definitive.” Another molecular biologist, Gunther Stent, said that “it really tells the whole story.” Stahl noted, many years later, that he had “been trying to do something half as pretty ever since.”7

In the same way that the form of “Easter Wings” reinforces its meaning, the tight, self-contained logic and cohesiveness of the Meselson-Stahl experiment strengthens its core message. Structure is a little more difficult to illustrate in an experiment than a concrete poem, so I’ll start with a reductionist approach.

Suppose we strip the experiment down to its minimal necessary information: the question (How does DNA replicate?) and the answer (Semiconservatively). I would argue that these are bare facts; on their own, they would probably not be considered beautiful. Let’s add in the hypothesis, which gives details on the copying mechanism (the DNA helix unwinds, and a new DNA strand is copied from each original strand. New DNA is thus made of one strand of old DNA and one strand of new DNA). Now the answer makes a little more sense. The explanation is logical, and it matches the pairing of bases in DNA. Next, we add back the experiment – how you go about testing this hypothesis (start only with heavy DNA and make new, light DNA from it. Then track how much of each type of DNA, heavy and light, there is over time). Under different hypotheses of DNA replication (semiconservative, conservative, or dispersive), we expect a different result, so this single experiment will immediately let us rule out two of the three. Finally, we add back the results (from heavy DNA, we go to DNA of intermediate weight, then a mixture of intermediate-weight and light-weight DNA). One hypothesis (semiconservative replication) is supported; the others are rejected.

By breaking the experiment down in this way, we can begin to understand how the individual pieces of the Meselson-Stahl experiment – question, hypotheses, experiment, and results – work together to form a coherent whole. First, every piece centers on a single issue (how DNA is copied), giving internal cohesion. At the same time, each new piece gives us additional information, providing the impression of direction. Finally, there is closure and completeness: The possibilities suggested by the three hypotheses are neatly tied up by the results. By the end of the experiment, we have come full circle with the answer to our question. The pieces of the Meselson-Stahl experiment strengthen and resonate with each other, and we perceive this resonance as beautiful.

Structure is important to a work in any genre – science or art – because it allows us to organize our understanding of the work’s messages and themes. When the structure focuses around and reflects the themes in some way, it provides additional power to the work. In this way, a close examination of structure can give us another lens through which to evaluate “beauty” across disciplines.


1 The same argument works on a theological level as well. The title, “Easter Wings,” tells us that this is a poem in celebration of Easter, or the resurrection of Christ after his death. In a reading from this perspective, short lines represent death, while long lines represent life or resurrection. The poet’s spiritual death and rebirth parallel the death and rebirth of Christ.

2 Six of the ten “-ame” and “-ee” rhymes even involve repetition of entire words: “became,” “me,” and “thee.”

3 Watson, J. and Crick, F. 1953. Molecular structure of nucleic acids. Nature 171:737-738.

4 Credit: DNA_simple.svg by user Forluvoft, Wikimedia Commons.

5 Credit: DNAreplicationModes.png by Mike Jones, Wikimedia Commons.

6 Meselson, M. and Stahl, F.J. 1958. The replication of DNA in Escherichia coli. PNAS 44:671-682.

7 Quotations from Holmes, F.J. 1996. Beautiful Experiments in the Life Sciences. In Tauber, A.I. (ed.) The Elusive Synthesis: Aesthetics and Science. Kluwer Academic Publishers, Dordrecht, pp. 83-101.

This entry was posted in BEACON Researchers at Work and tagged , , . Bookmark the permalink.

Comments are closed.