Visual identity in science illustration

This post is based on a talk I gave at Imperial College London in 2013 for Artifact, a society run jointly with the Royal College of Art.

What is science illustration?

I frequently get asked the question 'so you're a science illustrator, what do you do?'. Most people think of science illustration as designing illustrations for science textbooks. This is one of the things we do. We also try to represent research in a visually appealing way, either as info graphics or illustrations of things that aren't visible. Finally there's editorial illustrations where you have a lot of creative flexibility but don't need as much understanding of the technical details of the science.

At its heart all illustration is problem solving, trying to find the best visual solution to a set of constraints, called the brief. Some briefs requite absolute attention to technical detail, others allow for much more creative flexibility. Most illustration briefs can be placed somewhere along a sliding scale between technical and creative.

The illustration spectrum

At the technical end would be illustrations such as the engineering drawings in car manuals. There is no room for creative flexibility here because the primary purpose of them is to look exactly like the engine being worked on, perhaps in a cut-away form to help people navigate around them.

At the other end would be illustrations that required no scientific knowledge, just the appropriation of scientific imagery in a creative visual format. Examples would be ball and stick atom drawings or microscopes which were not there to illustrate the scientific content of an article, just to suggest that it fell into the science category. I have been informed that beauty products often do this in their advertising to appear to have more technical substance.

As is common with many fields, the interesting stuff happens somewhere in the middle of this scale. These are the illustrations that require knowledge of scientific topics in order to communicate the content accurately to the reader, yet still need to appear visually interesting. They will often appear in situations where they may be seen both by scientists and non-scientists. Scientifically literate readers will notice any scientific errors in the illustration, and treat associated material with increased scepticism. Equally they much appear visually appealing enough to draw the interest of non-scientists.

So if scientific illustration needs to visually appeal to non-scientist audiences, why does it often look so similar in style compared to traditional illustration?

The evidence

A quick google images search gives you a pretty good idea of the different styles prevalent in scientific vs traditional illustration.

The majority of science illustration images are dominated by a highly detailed pencil sketch style.

There is a much greater variety of styles in a search for illustration.

A search for another topic similarly constrained such as fashion illustration still produces a greater variety of styles than science illustration.

A similar theme appears when comparing the websites of illustration agencies for science illustration and traditional illustration. A traditional illustration agency will ensure that their illustrators have extremely contrasting styles, and that the home page of the agency displays these styles prominently. Indeed many agencies will refuse to take on illustrators whose style is too similar to an illustrator they already represent.

Science illustration agencies often make no effort to place contrasting styles next to each other, and sometimes do not even show images, listing only a directory of names and addresses, as if finding a science illustrators were as interchangeable as plumbers or cleaners.

A lack of demand?

For those who haven't guessed, I work as a freelance science illustrator. Occasionally we publish blog posts (like this one) or do interviews to get more exposure. The main point of this is to get potential new clients to contact us rather than having to chase them down ourselves.

I once received an email from a science textbook editor who had seen my illustrations in a blog post that I had written for a website. It went something like the following (italics are my reaction.

I saw your work online

Yes! I knew there was a point to writing long articles

I really like your illustrations

They like the artwork!

Your style is very unusual

I knew finding a niche and developing a style was a good idea!

Do you have any examples of more traditional science illustration?

…

I sent over a link to the rest of the illustrations I had on my website, but never heard back. I wrote it off as just having a style that wasn't compatible with what they were looking for. Later on I thought about it a bit more. There were certainly examples of infographics work in my portfolio that showed I had the problem-solving skills for textbook diagrams, and with a science degree under my belt the academic side of things wasn't the issue.

Perhaps it was just that my illustration style wasn't what they were looking for. But if it answered the brief, why was that an issue?

How did we get here?

There may be a good reason that most science illustration looks similar. After all, if illustration is problem solving, there's no reason why the best solution to a single problem shouldn't be the best solution to multiple similar problems.

Science illustration was originally to replicate things that needed to be observationally verified. Before the days of photography, you wanted to make sure that the red bird that you saw in the jungle was the same red bird that had been described to you in a lecture back in University. The easiest way to objectively verify this was to make an illustration of said bird that you could check it against.

Equally, when a surgeon was learning how to slice up a patient in order not to sever an artery, it was useful to have a visual reference as to where these arteries were to consult before or during surger

y. There was a limited amount of time and dead bodies that students could practice on, and none of them had blood pumping through their arteries to give you immediate feedback that you'd just made a grave error.

In both of these cases the illustrations were being used to compare against something that was observable without any specialist machinery. It therefore makes sense for them to be as true to life and photographic as possible to avoid any ambiguity. In both fields, this type of illustration is still useful due to the constraints of photographing human remains, and the difficulty in getting animals to pose for the camera.

Science however, has since expanded beyond these disciplines. Most cutting-edge research is not in the structure of the various blood vessels of the body (we think we've cracked that one) but in things that cannot easily be seen, such as proteins and gene interactions. Most scientists working with these will be more familiar with their three letter acronyms than the structures of said proteins. Photorealism is now not necessarily the best solution.

Where are we going?

Another new trend in science illustration has been a shift towards 3D rendering rather than hand-drawn illustration. This reflects the increase in computer processing power that allows complex renders to be made. It also allows the same 3D object files to be reused many times, many of which are easily available online. This means illustrators already know that their images will be anatomically accurate, regardless of the render style, and can instead concentrate on how best to highlight the important information within the structure. 3D has become particularly popular as once the structure of an object is pinned down, it is very easy to manipulate it and create a new render should your client decide on a different direction. 3D structures are also very easy to animate, even if this just involves moving a camera around the structure to get a better view of the 3D detail.

3D renders are also easy for scientists working in cell biology to interpret. They often mimic the structures visible down a scanning electron microscope, which is often the only way that cell structures can be made visible at high resolution and in a way that easily distinguishes them from other cells.

The appeal of the familiar

When creating an illustration of a subject that the viewer has no prior knowledge of, it is best to stick to photorealism rather than to make stylistic interpretations. You viewer will by default interpret illustrations as whatever they resemble in their real life experience. If you create a photorealistic rendering of a cell, the size of the cell and other physical attributes may be ambiguous due to the lack of real-world things to compare it to in the illustration, but the general structure will be obvious.

With a stylised illustration, such as a line drawing or flat diagram, the identity of each of the elements in the illustration is more ambiguous. In a diagram of a eukaryotic cell could look very similar to a town map when shown to a child with no idea what cells are. In this case the analogy has been used to explain the different parts of the cell according to how a child would see the town functioning.

This is important when illustration for children or students. For those familiar with the subject matter in question, or at least a vague background in the area, this kind of explicit referencing is not essential.

Interpretation of style

It is quite popular in the illustration world to draw a famous character in the style of several different illustrators. The point of this being that the essential visual cues that define a character are independent of the style used to represent them. This is a good method for illustrators and character designers to improve their skills as it helps them identify which features of a character are essential and should be kept, and which can be discarded. If a character is well known, viewers of the image will be able to recognise what they find familiar and draw parallels. These lego adverts are a great illustration of the idea.

This also works in science illustration. Linked are three different representations of eukaryotic cells created in the form of cake, cross-stitch, and crochet. In each of these cases a biologist familiar with the subject matter would probably able to identify the nucleus, the mitochondria and the endoplasmic reticulum.

Redefining the brief

If we were to start from the beginning, and ignore all preconceptions about what science illustration should and shouldn't look like, what would be the key elements that make up a good science illustration. In effect, what are the key constraints in our brief, and what elements can we have a little more artistic license with?

I can't answer this in an objective fashion as every brief will demand different things from both the illustrator and the viewer. I'm going to go through the key things that I think are important in science illustration, and therefore how these guide the work I produce. The relative important of different factors in an illustration will produce a slightly different brief, and therefore slightly different optimum styles. For instance I value simplicity over communicating volumes and textures. Someone who preferred the latter would probably represent science in a very different way to me. Neither of us are wrong, we just realise that in order to emphasise certain details of a physical system or structure, other details have to be omitted.

Technical accuracy

I think few would disagree with me that technical accuracy is extremely important in a science illustration. If you are asked to create an illustration to accompany a research paper, read the paper, and the accompanying material, then talk to the researcher about the material to make sure you understand it. Don't rush into creating an illustration based on what the author thinks would be the best way to represent the science. You will invariably make some kind of technical error in the way that you have depicted the science behind the illustration. If your illustration is published alongside the article in a journal, the readers may be confused as to why the illustration does not fit with the associated paper's description of a mechanism. You do not need to replicate every last technical detail, but you need to have a good reason for drawing something that is just plain wrong.

Examples include drawing the DNA helix spiralling the wrong way, which has a habit of annoying science illustrators in a similar way that misplaced apostrophes annoy writers. Yes we were able to understand the message, but the way it was presented implied that the person delivering it was a bit lazy. The structure of DNA is integral to the function of many DNA binding proteins, since they are specifically shaped to fit into the grooves. If you are going to draw DNA and can't be bothered to check which way around it spirals, at least draw it in a way that makes the direction of the spiral ambiguous so that your illustration doesn't deliberately misinform thousands of students when it appears in a textbook.

In the above illustration I decided to make a structurally accurate drawing of the Influenza A virus. This is often depicted as something resembling a laundry ball in many pieces of science illustration. Most undergraduate medical students would immediately label this as incorrect since the projecting proteins on a flu virus are not uniform. The virus is coated in a mixture of four subunit neuraminidase and three subunit haemagglutinin proteins. This is important as different strains of flu virus are often classified based on which version of the haemagglutinin and neuraminidase proteins they display. For instance swine flu is H1N1 and the most notable form of avian flu was H5N1. These classifications are often mentioned in non-specialist media such as news reports and therefore the difference should be illustrated. This is before we consider the fact that these proteins are essential to the mechanism by which these viruses enter cells.

Look different

I'm now going to back track slightly on what I said earlier about accuracy. Whilst you should never deliberately represent something in the wrong way, drawing something exactly as it appears makes for a boring illustration. Many processes in biology are dynamic. A single snapshot in time will not fully represent the process, and unfortunately we do not have the luxury of moving images to show the entire thing. Often you will require a little bit of artistic license to display all the key players in a process, since they are probably never in the same cell compartment at the same time. Equally if all your drawings are to exact scale, such as a drawing of the solar system, you will do a very good job of showing just how much empty space there is between planets, but won't be able to say much about the planets because they're just so tiny.

This is also one of your key advantages over photography or motion graphics. You are able to put things together that a photographer would never be able to capture. If the illustration you have created could compositionally be recreated fairly easily using a microscope or a camera, you are wasting your time. Obviously you want your illustration to have some resemblance to the structure you are studying, but you probably need to use less than you think. This point ties in to what I said earlier about accuracy. It's OK to take artistic license with the science if and only if you have a good reason for doing so. Be smart, don't be lazy.

Above is an illustration I created for a seminar about centriole assembly. Centrioles have a distinctive 9-fold rotational symmetry. I was considering simply illustrating a cross-sectional view of this, but thought that this was inappropriate for a dynamic mechanism such as assembly. In order to make the illustration recognisable as a centriole, I left in just enough of the 9-fold symmetry skeleton as I could, whilst using each of the repeating segments to represent a separate step in the assembly. The mechanisms was fairly well detailed in the accompanying paper, although how the second and third microtubules attached was still an unknown.

Simplify the information

When I'm not drawing science I also work as a graphic designer, and one of the key principles to effective communication design is to not include anything that may distract the viewer from the things they're meant to be focusing on.

Back when I was a finalist at university and had no idea what to do with my life, I applied to one of the world-leading branding agencies to be an intern. The head of HR organised an interview for me with the Creative Director for Europe. I brought along my portfolio and went through it for half an hour whilst he listened without saying a word. Once I'd finished, he flicked back to a piece of graphic design that I'd produced for a student society and pointed out that I'd placed a colour change in the middle of the page at the same time as changing the font size. I had given my viewer two different pieces of information to signify a single change in hierarchy. Not only that, I had added a border around the letters in the title, this adding extra lines to the composition and distorting the letterforms. He then proceeded to point out similar flaws in the majority of the rest of the text-based work that I produced (not a great start if you're looking to work at a company that designs logos).

The general lesson I learnt that days has stuck though. Every extra line or colour change you add to a composition is extra information. If that information is not necessary, or does not signify something important, it is a distraction from the information that does. Do not add extra colours or lines unless they are absolutely necessary to understand the information correctly.

When you create a composition, you have to work out which piece of information is the most important piece to communicate. Your choice will influence how effectively you can communicate the rest of the information, so it is important to isolate the most important and concentrate on that, bringing it out with bright colours. If you genuinely have two different pieces of information that are equally important, do two drawings. It is better to communicate one thing well than two things poorly. Once you have worked out what the most important thing to communicate is, work out everything else that it is necessary to include in order to communicate the key thing well. You should still include this information, however it is of secondary importance, and therefore should use more muted colour schemes. Any information that does not fall into one of these two categories should be discarded.

In the above illustration I was trying to make a more effective diagram of how the nerves and arteries in the hand are arranged. Most anatomy textbooks feature illustrations which, whilst technically accurate, overwhelm the viewer with detail and require an extensive amount of time to decipher. The key things I wanted to illustrate were the nerves and arteries so I represented these in the highest contrast colours available. Originally I was going to simply feature the silhouette of the hand, however I felt that this lacked the context of how the veins and arteries were arranged with respect to the bones. In the hand the bones are the most obvious things to find, and the palm of the hand is quite deep. I felt that it was necessary to include the bones in a subtle way to show that the hand was palm-up and that the majority of the arteries did lie on that side of the hand (which is obvious if you look at your wrist).

Examples

Below are links to some examples of science illustration that I think follows the key principles listed above, simplicity, technical accuracy, and looking different.Human Body by Clear as Mud at www.clear-as-mud.co.uk

This human body illustration by Clear as Mud provides a simple way to illustrate how the human body internal organs and skeletons are arranged. A few liberties have been taken with the intestines, which don't make the distinction between the small and large intestine, however this is an understandable compromise to make the shape of the pelvis more visible. I personally think that this illustration would be very well-suited to a KS2/KS3 science textbook as students at this level often have to draw the human body, and a simplified diagram such as this is much easier to replicate in the exam that the ones often shown in textbooks.

This illustration by Alison Haigh is a clever and very simple representation of the atoms in the periodic table looking at just the number of electrons in their orbitals. Haigh has clearly researched her subject matter carefully, correctly showing that certain D-block elements (such as Chromium and Copper) preferentially fill their 3d orbitals over the 4s orbital. My only criticism is that by separating electrons into the main energy levels, it makes it less clear when an atom has a full outer shell or an almost-filled one (see groups 7 and 8).

Summary

In conclusion, different styles of illustration suit different areas of science, due to the differing nature of their briefs. It is a mistake to assume that just because a certain style is seen as 'the norm', it should be used for everything. Branch out, try something new!

All illustrations (c) their respective artists

January 28, 2020

claudiastocker