“Data of the most important kind”

Matthew Cunningham
Sunday 25 August 2019

I’ve had nearly a month now to reflect on my first five weeks of research and appreciate the winding dusky forest path that academic research takes. I am undertaking a History of Science project with a background in Pure Mathematics and no past experience in academic History at all. My research into the increasing development and use of experimental curve plotting in the mid 1800s has therefore been entirely new land for me to explore. As well as an excellent opportunity to try a new discipline and practise different academic skills.

Curve plotting is a way of visually representing collected data by plotting points on an axis and then putting a ‘line of best fit’ through those points to show the trend they follow. In science this is now an indispensable way of analysing and presenting results of experiments. It shows clear relationships between different variables which could not be seen from raw data and, as I shall discuss later, allows for the easy physical interpretation of numbers. However, before the mid 19th century it was a rare, eccentric, esoteric method with very few practitioners. My first week of research was spent burying deep into the secondary literature (not simply to learn about my area of history but also to understand the aims, methods and style of academic History papers). Much of this was focused on the use of curve plotting, but I also studied writing on the history of changes in representation culture, the state of science education in the UK at that time and, of course, biographies of the scientists whose work I was interested in. All these topics were recommended to me by my supervisor and it quickly became clear how specialist this discipline really is, and how novice I really was. From my reading, I learned that by the 18th century the mathematical tools for graphs were well understood but they were still rare and unfashionable; with the exception of J.H.Lambert (who produced many a sophisticated graph) and a follower of his. Despite Lambert’s successful use of graphs the method did not catch on and disappeared after his death. By the 1830s, graphs were starting to appear again in the work of British scientists. There are many reasons posited for the explosion of use of graphs for analysis in science at this time which need investigating. Some of these are the availability of graph paper, the use of automatic recording instruments and a change in visual culture in science (and society as a whole). Automatic recording instruments (machines which note experimental results without a person present) are particularly interesting because, due to their nature, they often output a curve of sorts. However, the scientists would usually then look at the curves and pick particular data points to read off and make a table from them – such was the culture!

My own research for my first summer focussed on J.D.Forbes and his part in the progress of curve plotting. He was Professor of natural philosophy at Edinburgh University for most of his career and one of the most important scientists in Scotland at the time. His papers are in the St Andrews archives and so he is the focus of much History of Science research here. In my second week of the project I started looking at the vast Forbes collection. I discovered that there is much diversity in the kinds of documents we have to study an academic like Forbes. He wrote travel and research journals, scientific papers, articles, lecture notes, and thousands upon thousands of letters. Knowing how to find what you’re looking for is not an innate skill and must be learned by extensive experience. I looked at thousands of pages of letters without being able to read them particularly well (handwriting is another difficulty with History of Science research and complaining about it is the primary way to break the ice when meeting fellow academics) and found very little at all of use. I also studied the syllabuses he wrote and many of his journals, only finding hints of what I was looking for. These are significant findings of course. The fact that he didn’t spontaneously draw curves during field work or have curves as a topic of his teachings shows that they weren’t as essential in experimentation as they are nowadays. As good as a discovery of absence is, it was disappointing and demoralising for me at the time. It wasn’t until my 4th week of the project that I started looking in the right part of the collection for my findings. From my discoveries here I’m going to relate some of the ones I found from his  work on Meteorology, since I found them particularly exciting. Indeed, in what I believe are a set of lecture notes, Forbes writes, ‘The graphical mode of representing facts early applied to meteorology.’ This sentence is somewhat ambiguous because, although it likely means that graphical representation was used early in the development of meteorology, it could potentially be a way of saying that one of the first uses of the graphical mode of representing facts was in meteorology. Forbes’ papers on meteorology make use of and discuss many different curves. For example: he suggests a method for calibrating a thermometer using an ‘interpolating curve’ through data of the difference between the new thermometer and one known to be correct against the temperature shown. He looks at two curves of temperatures in different places and states that they are similar except for in three ways and goes on to analyse these differences and their physical implications. He explains how you can change the temperature scale you are looking at by moving a curve up and down on its axis (and he explained this as if it would be novel to his readers – showing how unfamiliar the scientific community was with experimental graphs). Forbes sums up why he has started using experimental curves in a rather beautiful quote: “The examination of these curves furnishes us with some data of the most important kind for Meteorology which it is best in the first place to state, and afterwards to consider how we can explain.” His stating so carefully that such curves are to be used to analyse results from an experiment to find trends and peculiarities that need to be explained clearly shows that curves were novel to his audience. He is also giving the method his highest possible praise, which would have made a strong impression on his audience in view of his important position in Scottish science and the large number of correspondences he had with the most prominent leaders of the scientific community in his day. 

These findings are compelling evidence that Forbes was part of the culture change that lead to curve plotting’s central position in our scientific handbook. Next summer I would like to study this shift further by investigating lecture notes and student notebooks from the mid 19th century from other Scottish universities as well as the work being done during this time by other scientists who have been identified as being early adopters of graphs. The educational sources are particularly interesting to me because when a method is being taught generally it must have a wide acceptance in the academic community. Therefore it would be an excellent indicator of when the use of graph became normalised. I will also be studying the secondary literature more broadly to gain enough knowledge of this time and the practices of academics in this field, with a view of writing a publishable article on my findings.

I would like to give many thanks to my supervisor, Isobel Falconer, for her amazing insight and help during these first 5 weeks of my project and I look forward to next summer.

[1] Tilling, L. (1975). Early Experimental Graphs. The British Journal for the History of Science, 8(3), 193–213. https://doi.org/10.1017/S0007087400014229

[2] Hankins, T. L. (2006). A “Large and Graceful Sinuosity”: John Herschel’s Graphical Method. Isis, 97(4), 605–633. https://doi.org/10.1086/509946

[3] Forbes, James David (1809–1868), physicist and geologist | Oxford Dictionary of National Biography. (n.d.). https://doi.org/10.1093/ref:odnb/9832

[4] Phases of physics in J. D. Forbes’ Dissertation Sixth for the Encyclopaedia Britannica (1856) – Isobel Falconer, 2018. (n.d.). Retrieved June 18, 2019, from https://journals.sagepub.com/doi/10.1177/0073275318811443

[5] St Andrews Archives: Special collections. Books with class marks : MS Q113.F88, MS Q113.F82, MS Q113.F85, MS Q113.F89.

Share this story

Leave a reply

You must be logged in to post a comment.