Watson and Crick's model of DNA, 1953. Credit: SCIENCE MUSEUM/SCIENCE & SOCIETY PIC LIB

When Linus Pauling and Roger Hayward called their 1964 book The Architecture of Molecules they were resorting to an age-old style of metaphorical allusion. As knowledge of molecular structure has entered worlds that are impossible to ‘see’ in any normal sense, analogies with objects in our normal range of visual experience have lost none of their efficacy as aids to visualization.

One of the most spectacular acts of modelling was initiated in 1937 by Max Perutz's work on haemoglobin. As Perutz graphically recalled, “the first protein structures revealed wonderful new faces of nature”. That he found it necessary to reveal these “faces” visually rather than mathematically became rapidly apparent: “I began giving a course in X-ray crystallography of biologically important molecules for students of biochemistry and other biomedical subjects. In my first lecture I introduced lattice theory, trigonometric functions, and Fourier series⃛ but half the students failed to turn up for my second lecture.⃛ The following year I replaced my forbidding lecture with a non-mathematical, largely pictorial introduction called ‘Diffraction Without Tears’.”

The potency of the visual model goes far beyond the need to sweeten the mathematical pill for students. The three-dimensional model has proved a vital tool in showing how new substances can be “engineered”.

A.F. Cullis et al.'s model of the haemoglobin molecule, from Max Perutz's Proteins and Nucleic Acids: Structure and Function, 1962. Credit: ELSEVIER

General agreement on a definitive model to ‘fit’ a particular X-ray diffraction pattern — as happened when James Watson and Francis Crick's model of the structure of DNA crystallized Rosalind Franklin's experimental data — does not solve the fundamental problems of what sort of model is appropriate. The early range extended from the spidery lattice that Watson and Crick revealed to the world in 1953 to the weighty sculptural edifice of the 1962 haemoglobin model.

What is striking to a historian of visual images is how the models have period ‘styles’. The compound of factors that make up this style — the ‘look’ or visual ‘feel’ of the object — include materials, constructional techniques, colours, textures, scales and the vocabulary of the shapes. The choices, in science no less than design, involve complex permutations of utility, technology and (often inadvertently) aesthetics.

Our sense of the period style of the Watson-Crick model — linear, wiry, openly mechanical, unadorned and rhetorically ‘functional’ — is framed by reference to earlier and later systems of representation. Their precarious spatial lattice stands very much within the design parameters of the 1951 Festival of Britain, the event that marked the British embrace of a modern style for a new age. In contrast, the ‘Glyptic Formula’ kits for modelling molecules in the nineteenth century, with their polished balls, firm rods and turned mahogany stands, exude the air of a gentleman's billiard room. And recent computer images parade the high-tech rhetoric of electronic graphics.

The haemoglobin model has its own 1960s look — assertive and futuristic like a visionary model for a concrete block of layered residences. As in any work of architecture, more is involved than mere structure.