At an event celebrating the opening of a new University of Toronto physics building in 1907, University President James Loudon (1841-1916) described the challenges that he had faced in setting up the University’s first physics laboratory in the late 1870s. Among other topics, he spoke about the importance of craft skills to furnishing basic laboratory apparatus.
In one of my first difficulties—how to make air-tight joints between glass and iron tubes—I applied to the late Professor Rowland, of Baltimore, and received the advice to break a dozen tubes in discovering a way. To learn by making mistakes is not a bad plan, if you can afford it. [Loudon 1907, 45]
Loudon’s anecdote exposes a facet of scientific practice that we don’t often consider. Certain laboratory tasks require researchers to master difficult manual skills. This poses a challenge to scientific communication because such tasks (as implied in the advice given to Loudon) are difficult to teach and to learn. We can explore these processes by “re-enacting” or “recreating” them—an approach that has been developed by historians of science to expose those elements of an experiment that remain unreported.
For the past year or so, I have been studying the process of making glass micropipettes that, beginning in the first decades of the twentieth century, have been used to inject material into, or remove material from, individual cells. This is part of a bigger project to remake various missing parts from a Chambers’ Micromanipulator that was acquired by the University of Toronto in the late 1920s or early 1930s. The experience of remaking these delicate glass objects could be taken as an example of how we might study the craft-like laboratory skills that used to (and still do) underlie various aspects of scientific research, especially in the early stages of a technology’s development before novel tasks are automated and become “black boxed.”
The type of glass microsyringe that I am seeking to recreate would have been mounted in a microinjection apparatus used with a Chamber’s micromanipulator. This system is depicted here with the glass microsyringe (circled in green) mounted in the removable tip of a tool holder.
By the late 1920s, nearly every part needed to operate a micromanipulator could be purchased from a Leitz catalogue. Glass microsyringes were an exception because they were very fragile, easily clogged, and had to be clean and sterile. They were considered disposable and were typically made just before they were used.
As these tools became common in medical and biological research laboratories over the first half of the twentieth century, many researchers gave their own instructions for making them when describing the apparatus used in their experiments. The illustrations on the right accompany a particularly detailed set of instructions given by Robert Chambers (1881-1957) , inventor of the Chambers’ micromanipulator and, for decades, the most prominent researcher in the field of micrurgy.
As noted earlier, this process involves softening a glass tube of 3-5mm in diameter in the flame of a Bunsen burner and pulling it into a capillary of between .5 and .85 mm in diameter. This step is not shown here, probably because the instrument’s users would have known how to “pull pipettes”.
The photos show the making of a tool tip using the pinhead-sized flame of a microburner to soften a section of this capillary while applying tension to both sides of the tube. Done properly, the capillary separates into tapering tips. The base of the tip is then bent upwards in the heat of the microburner flame and the finished micropipette is stored for later use. When needed, the capillary is shortened to about 5cm and placed in a tool holder to be mounted in the micromanipulator.
Chambers’ description of the critical moment in which the tip is made conveys the challenge of this process:
The capillary will separate with a slight tug—a feeling much like that experienced when a taught thread, held in the fingers, is parted in a small flame… Everything depends upon the amount of heat used and the timing of the added pull, and these vary slightly with the height of the flame and the diameter of the capillary. With a little experience, one can usually tell when a proper tip is made by the peculiar feeling just described. [Chambers 1951, 81]
Even when successful, certain tips were more suitable to a particular research task than others. Inspection under a microscope would reveal exactly what kind of tip had been created. Both the challenge and the variability of this process led researchers to develop instruments to automate and regulate it.Remaking: setup and supplies
Historians usually take on these sorts of projects with help from museum conservation departments, university technicians, or backing from other well-funded institutions. For various reasons, my work with the Chamber’s micromanipulator is both and academic collaboration and a DIY project. While I have had generous help on various aspects of the project from the machine shop at the U of T Department of Physics and the Semaphore maker lab at the Department of Information Studies, I have made the glassware on my own.
The most basic laboratory glasswork requires a supply of natural gas. (Serious laboratory glasswork, generally done by professional glassblowers, requires both compressed propane and oxygen.) Since working scientists tend avoid letting historians play with fire in their laboratories, I have had to improvise. The straight forward solution would have been to purchase a Bunsen burner, a small propane cylinder, a low pressure gas regulator, and a length of rubber tubing. I settled on a cheaper solution: a $7 camping stove that mounts on a $5 bottle of propane/ butane camping fuel. With a small central burner, this stove is very serviceable for softening a Pyrex pipette, though probably less efficient than a stove with a wider ring-shaped burner for heating food in the wilderness.The microburner was more of a challenge. Historically, this was usually just a narrowed piece of hard glass or a hypodermic needles attached to a gas supply with a rubber tube. The gas was regulated by placing a pinch valve over the tube. After some tinkering, I ended up with a wooden mount for a Bic lighter. The mount features a swiveling arm that can depress the valve just enough to produce a tiny flame.
I was very lucky in my search for suitable glass tubing. Active Surplus, a maker’s paradise in downtown Toronto, sells surplus capillary tubing of roughly the right size. At 2.9mm in diameter, these are very close to Chambers’ recommended range of 3 – 5mm. At 50 cents for a tube 9.5cm long, this was much cheaper than ordering from an online supplier.
Playing with fire
Using this equipment, I was able to follow the Chambers’ instructions for making micropipettes. I found the process of producing the finished tip to be just as finicky a Chambers described, though the process of consistently pulling the 2.9 mm tubes into capillaries of .5 – .9mm in diameter and 5.5 – 8cm in length was much harder than I expected. The final step of bending the tip is especially frustrating because it is very easy to melt the tip itself.
There are several factors that need to be balanced throughout: the intensity of the heat source, the amount of material to be heated, and the degree of softness that the material reaches. It is easy to overheat the glass and produce a long flexible filament.
Making these tools reminded me of the frustrating process of learning to serve a squash ball. In both cases, I found myself trying to reason my way through by constantly adjusting my approach, but to no real effect. The advice to “break a dozen tubes in discovering a way”—akin to “keep playing squash until you can serve the ball”—seems to apply.
Craft skill and “gestural knowledge”
In studying the production of scientific knowledge, historians have begun to notice the “non-verbal, non-articulate and preconceptual elements of experimental activity.” [Voskulh 1997, 338] This emphasises the role of the experimenter’s skill in a process that tends to be conveyed, whether through text or demonstration, in such a way as to stress the self-evidence of a particular finding. A masterful article from 1995 by Hanz Otto Sibum introduced the notion of “gestural knowledge” to encompass the “complex of skills and forms of mastery” necessary to produce an experimental effect, though he also applied the term to craft skills more generally. These two senses intersect in his exploration of James Prescott Joule’s (1818-1889) experiments on the mechanical equivalent of heat. [Sibum 1995, 73-74]
Sibum interpreted Joule’s skill with the thermometer against the backdrop of the Manchester brewing industry of the mid-19th century where practical, imitative knowledge was being supplanted by the highly precise methods of practical chemists. Joule, the son of prosperous brewers, depended on this tradition of monitoring the brewing process using thermometers for his success as an experimenter. Sibum’s insight came, in part, from his own attempt to re-enact Joule’s experiments using recreated apparatus. [Sibum 1995, 83-95]
The process of making glass microtools is, in certain respects, similar to the traditional skills of the malsters and brewers or Joule’s modern skill with the thermometer. All are forms of manual mastery obtained through experience. All represent the intersection of craft skills with laboratory research.
These examples differ, perhaps, in the sense that historians of science tend to invoke craft skill in search of the hidden and implicit within an experimental account—Sibum’s re-enactment of Joule’s experiment demonstrated the care and skill needed to obtain precise results. Rather than eliding the difficulty of a scientific operation, Chamber’s instructions are explicit about the manual challenges involved. Tracing similar instructions for making glass microtools over the history of micrurgy, one sees them becoming more detailed with time as the difficulty of the gestural knowledge required for the task became more evident among a widening pool of new users.
In fact, it is a bit misleading to compare this rudimentary laboratory process with the sorts of experimental claims that historians have tended to study through re-enactment. For this reason, I’ve come to prefer the term “tactile history of science” to others such as “experimental re-enactment” or “scientific recreation” because it broadens the scope of the remaking/ re-enactment process beyond the experimental report to include the everyday operations of laboratory work. I’m not sure where the term “tactile history of science” originated, but I encountered it on Will Thomas’s blog, Ether Wave Propaganda, where you can find a good explanation and an excellent series of posts on the topic. Compared to terms that emphasize experimental claims, I think that it gives historians a broader mandate to study the scientific researcher as a maker while giving makers an entrypoint into the history of science.