Content Server

TEST OBJECTS FOR MICROSCOPES

Jutta SchickoreIndiana University Bloomington

INTRODUCTION: LONG-TERM AND SHORT-TERM HISTORY OF TEST OBJECTS

This article examines test objects for microscopes: specic objects designated to test the quality and performance of these instruments.1 I trace the history of these objects to understand the development of microscopy as a research technique and the methodological concerns of its practitioners. Test objects can be historicized in two ways. First, test objects themselves have a history: they emerged in particular social and scientic contexts; the contexts of application multiplied; and the requirements for good test objects changed over time. Secondly, test objects are part of a long-term history of evaluation criteria for microscopes. Since the seventeenth century, practitioners have assessed the quality and performance of their instruments; but over the centuries the evaluation criteria and procedures have changed. In what follows, I consider both the short-term history of test objects and their role in the long-term development of microscopy as a research technique.

My article extends the scope of Graeme Goodays study of late Victorian micro-scopy.2 Gooday has described the practical problems experimental physicists and biologists faced in their attempts to get their devices to work reliably. He shows that the microscopists turned their attention to the effective management of the physical environment in which they practised their research, seeking to obtain stability in the optical instruments as well as in the investigative setting. While the practitioners agreed that successful microscopical observation required the effective management of working environments and that illumination and lens construction were the key factors in microscopical observations, they disagreed as to what constituted the best ways and means of management.3 Gooday also draws attention to the practitioners reports of the disturbances, interference, and impediments they had encountered in their work. He argues that these reports were crucial to establish the credibility of the practitioners.

Like Gooday, I examine the practitioners efforts to stabilize the conditions of micro-scopical observation and, in particular, of reliable testing. But my historiographical perspective is different in that I put the debates and practices surrounding test objects centre stage. Gooday concludes that in the longer term, microscopists succeeded in stabilizing their working environment. But as I show in this essay, the practice of using test objects never became fully stabilized. Tracing the history of the test objects through the second half of the nineteenth century, one nds that the criteria for good test objects in fact became more and more problematic as they were further specied. Moreover, it turned out that these requirements could never be fully met.

My essay has four parts. The rst three sections examine the short-term history

0073-2753/09/4702-0117/$10.00 2009 Science History Publications Ltd

Hist. Sci., xlvii (2009)

118 JUTTA SCHICKORE

of test objects. I begin with a brief sketch of the intellectual and social contexts in which test objects were introduced and trace their swift establishment as key tools of microscopic practice. I show that in the second third of the nineteenth century, the contexts for testing multiplied, and test objects came to be used as reference points to draw boundaries between different groups of practitioners: the microscope enthusiast, the (mere) practitioner, and the (real) man of science. The second part traces various aspects of the career of test objects. I show how the preoccupation with tests and testing posed several new challenges to the practitioners and even had destabilizing effects. Uniformity, familiarity, and difculty of test objects emerged as crucial requirements for successful testing. But the practitioners continuing strug-gles to identify good test objects that could meet these criteria and to specify ideal testing situations indicate that these requirements could not be fully met and indeed that it was not even clear what they meant.4 In the third part I draw out the dynamics of the career of test objects and show that the practice of using tests was never fully stabilized. Close to the limits of resolution, there was always a trade-off between epistemic gain and risk.

The concluding section characterizes the role test objects played in the long-term history of criteria and procedures for quality checks for microscopes. Seventeenth- and eighteenth-century microscopists too assessed their instruments; but over the centuries the evaluation criteria and procedures changed. The introduction of test objects profoundly transformed the ways in which microscopes and microscopical observations were evaluated. Test objects made the practitioners realize that micro-scopes had to be assessed on a large number of different properties, and that it was impossible to make a microscope that was perfect in all respects.

ASTRONOMY, NATURAL HISTORY, AND THE MULTIPLICATION OF CONTEXTS FOR TESTING

Test objects emerged in early nineteenth-century Britain in a social context that was conducive to comparison and competition. Microscope enthusiasts scientic gentlemen, medical doctors, and instrument makers often came together to study curious microscopic objects. Instruments were brought to these informal gatherings, their mechanical and optical properties were compared, and people discussed how well they fared.

In the early decades of the nineteenth century, different types of microscopes were available from a variety of makers. At that time, simple microscopes often outperformed compound microscopes, which suffered more from various kinds of aberrations. But simple microscopes were rather awkward to use and had a narrow eld of vision, and therefore many people were hoping to advance the optical per-formance of compound microscopes. One of them was the physician and microscope enthusiast Charles Goring. He occupied himself with natural history, had a keen interest in astronomy and optics, and kept various instrument makers busy with sug-gestions as to how to improve the performance of compound microscopes.

Gorings experiments on the instrument established that a range of material fac-tors conditioned the effectiveness of a compound microscope. In particular, he noted

TEST OBJECTS FOR MICROSCOPES 119

that one cause of the inferiority of compound microscopes was a small aperture. To account for this fact, Goring appropriated the term that had been introduced to characterize the optical performance of telescopes, namely, penetrating power. In 1800, the astronomer William Herschel had coined the term to refer to a telescopes power of making faint and distant stars visible, and he had used distant clusters of stars to ascertain the penetrating power of telescopes.5

Goring appropriated both the term and the procedure. Combining the concepts and practices of contemporaneous astronomy with natural history, he suggested using organic objects with ne lines and patterns as reference points to evaluate microscopes. In several publications from the late 1820s, Goring described and depicted diverse objects that were suitable for this purpose, such as insect wings and ies feet. Goring called such objects test objects. Echoing Herschel, he dened a test object as some object which can only be seen by a microscope single or compound, which possesses a great quantity of good distinct light (the result of a large and perfect aperture).6 Goring explicitly called the test object a standard of comparison, whereby to scruti-nize the pretensions, the defects, and excellencies of different microscopes, and prove beyond dispute whether they are effective or not.7 His co-worker, the instrument maker Andrew Pritchard, again made the connection to astronomy explicit. In an essay exclusively devoted to test objects, Pritchard noted that since the discovery of double stars and nebulae, telescopes had to undergo more severe tests, and greater accuracy in their construction is required. What has been advanced in regard to the telescope will be found applicable to the microscope.8

It is a little odd to ascribe penetrating power the telescopes power of reach-ing out into the depth of space to microscopes. Gorings optical theories were not well received. His appropriation of Herschels notion of penetrating power to microscopes sparked heated debates about whether it was correct to apply that term and what exactly penetrating power was. Goring soon began using denition or dening power, and the introduction of this term further complicated the debate. The discussions appear rather obscure to the twenty-rst-century reader because they pertain to the physical causes for certain optical properties of microscopes and at the same time to the appropriateness of the terms for those properties.

A few examples may illustrate the situation. In his comprehensive manual of microscopy, for instance, Pieter Harting foresaw great confusion over the term penetrating power because some practitioners had begun using it to denote a measure for the depth of the visual eld.9 Indeed, in his instruction manual The microscope and its revelations William Carpenter distinguished penetrating power or focal depth from both resolving and dening power.10 The botanists Carl Ngeli and Simon Schwendener also claimed that the microscopes so-called penetrating power was something entirely different than the telescopes penetrating power. However, unlike Carpenter, they noted: It is always the oblique rays, which penetrate and the central rays, which dene, but actually both do the same.11 They complained that the introduction of test objects had produced the unclear notions of penetrating and dening power in the rst place.12 Other practitioners took issue with Gorings

120 JUTTA SCHICKORE

ideas about the working of the microscope. They debated whether penetrating power was a defect or an advantage,13 whether a larger aperture was always advantageous,14 and so on. Even in the last decades of the nineteenth century, no agreement had been reached about these issues. But with the benet of hindsight one may say that the perhaps unjustied transfer of an astronomical concept to another context was pro-ductive despite or perhaps because of the confusion it caused, as these debates certainly contributed to the understanding of the microscopes working.

In contrast, the practice of using test objects was quickly adopted. Within a few years, all over Britain and the Continent nely-lined objects were used to evaluate the performance of the microscope. One can nd reports in journals of social gatherings where microscopes were tried with test objects.15 A couple of years after Gorings articles had appeared, the Viennese physicist and microscopist J. v. Jacquin praised Dr. Goring in London, who had been the rst to draw attention to test objects.16 Test objects were sent to France so that French microscopists might try their microscopes on them,17 and on the other side of the Atlantic, test objects were also in use. Instru-ment makers sold their microscopes together with sets of test objects.18 Anatomists began to tell their readers what test objects their microscopes were able to display, so as to prove that their observations were reliable. Instruction manuals and textbooks of microscopy included detailed sections on test objects, and entirely new contexts and goals for testing opened up. In the remaining part of this section, I characterize the main contexts and purposes of testing.

The initial, comparative use of tests continued to dominate the social gatherings of microscope enthusiasts. In the late 1820s, Thomas Gill, the editor of the journal Technical repository; or, discoveries and improvements in the useful arts, repeatedly printed reports of get-togethers at which different types of microscopes from various makers were put to test and compared. In 1827, the assembled gentlemen even had the opportunity to witness how well English microscopes fared in comparison with those produced by Giovanni Battista Amici. The renowned Italian instrument maker visited London at that time and had brought his microscopes along. He participated in the social gatherings of microscope enthusiasts and his microscopes were tried with various tests.19

The preoccupation with test objects and particularly the public trials and competi-tions stimulated exchanges anong practitioners, instrument makers, and microscope enthusiasts. Test objects thus became motors of technological advancement. This is a noteworthy point because many of these competitions were staged by microscope enthusiasts or what one today would call amateurs. These scientic gentlemen who were perhaps not immediately involved in scientic research were ideally placed to communicate with and challenge instrument makers to develop improved instruments. Gills reports show that the English instrument makers who were impressed by the performance of Amicis microscopes set out to utilize Amicis principles in their own workshops.20 Already in 1833, Pritchard attributed the grand and magnicent improvements which the microscope has recently received to the discovery of objects that may be considered as tests of the penetrating and dening powers of this


instrument.21 At that time, this may have been largely an expression of optimism, but the interactions between microscope enthusiast and instrument makers surely were incentives for the artisans.

Several practitioners explicitly mentioned that their engagement with particu-lar test objects had spurred their ambition and declared that the engagement with test objects was benecial to microscope making. Others even suggested specic directions of research into microscope technology.22 In 1870, some fty years after Goring and Pritchards rst articles, the British geologist James Scott Bowerbank reminisced about how he and his friends had carefully assessed every new improve-ment of lens combinations with test objects, thus having done their best to incite the makers to aspire to the greatest possible perfection in the construction of their object-glasses.23

There was an element of national rivalry often friendly, at times erce in the comparisons of microscopes, something one can also nd in contemporaneous glass and telescope making.24 Both Gill and the instrument maker Pritchard used test objects as vehicles to perform comparisons across the Channel. Gill reported that the eminent French opticians, Chevalier & Sons, of Paris had written to him and requested that a few test objects be sent over to them. Gill told his readers that he had been pleased to comply with their request and sent them various kinds, and among them those of the brassica and podura, desiring them to favour me with the result of their examination: and by this days post (August 5th), I received a letter from them complying with my request, and accompanied with drawings of the two scales, as they appeared under their microscope.25 Pritchard also noted that some of the test objects he and Goring had singled out were immediately transported, that our neighbours the French might try their microscopes on them.26

Test objects even travelled all the way to America.27 The letter written in the 1840s by the naturalist and professor at West Point, Jacob Bailey, to the instrument maker Charles Spencer illustrates both the interactions between instrument maker and prac-titioner and the amicable competition between the English and the Americans. Bailey reported how excited his London correspondent had been when he had received the latest American test objects. He and his friends had tried it immediately and could make nothing of it, even with the nest Smith, Ross, and Powell glasses. Now the Englishman was desperate to have an object glass made by Spencer and inquired about the costs. Bailey advised Spencer not to send anything over until you are sure of astonishing the natives.28

National rivalry in the instrument making business and the competition between English, American, and Continental makers continued throughout the century;29 and designated test objects provided the tools for the competition even across long dis-tances. In 1888, the American microscopist H. J. Detmers reported to the American Society of Microscopists his efforts to establish how the best American instruments compared with the best European ones. He submitted to the Society some photo-micrographs of the then common test object, the diatom Amphipleura pellucida, as well as of bacteria. He had made the images with a microscope objective produced

122 JUTTA SCHICKORE

by the American maker Robert Tolles. Detmers challenged those among his com-patriots who still did not trust American instruments to produce better photomicro-graphs with their European microscopes. On a trip to Europe, Detmers had visited three major German optical rms, Leitz and Seibert & Seibert in Wetzlar together with Zeiss in Jena. Detmers stated that all three opticians had been impressed by his photomicrographs and that especially the images of Amphipleura had been the best they had ever seen.30 Both Zeiss and Robert Koch famous for his work on bacteria and on microphotography had given him photographs of Amphipleura and of bacteria, which he submitted to the Society for inspection. Detmers pointed out that unlike the images he had produced, neither of the German photographs was free of diffraction lines. In conclusion, he stressed that he was more than ever convinced that the renowned German instruments did not in any way surpass the best American ones.31

In addition to the function of test objects as tools for comparison and competition, Gorings initial denition of test object emphasized the role of these objects as quality checks. By displaying the strengths and weaknesses, scope, and peculiarities of individual instruments, test objects served the practitioner to validate the microscope at hand and, vicariously, the observations made with it. As Goring characteristically put it, the test objects would give the user the power to examine the pretensions of all microscopes whatsoever, with infallible accuracy, and the unblushing effrontery of the shoptician will in vain attempt to palm worthless trash on the ignorance of the public.32 In a similar fashion, if with less effervescence, his co-worker Pritchard claimed that scales of cabbage butteries, the tips of tiny leaves, and other delicate objects offered the best means for the naturalist to determine the exact capabili-ties of his instrument, in order that he may not be led astray in his investigations, by placing undue condence in it.33

Pritchard made this statement in his essay on test objects. His main aim was to identify certain natural objects that could serve the individual practitioner to evaluate the powers of the microscope at hand. In passing, Pritchard mentioned another group of objects: articial stars, enamel dial-plate, wire gauze, &c.,34 which were also used to check the quality of the microscopic image.35 Goring used small black enamel plates with white numbers to check for spherical aberration.36 Other microscopists used different devices for similar purposes. The German microscopist Hermann Frey, following the botanist Hugo von Mohl, utilized a slide covered with India ink, in which small circles or gures were scratched.37 In contrast to the organic objects, these mechanically produced objects served a technical purpose. They were used to investigate the physical causes of imperfections of the microscopic image.

Pritchard did not go into further detail about this technical dimension of testing. But he did utilize his brief reference to hand-crafted or engineered test objects such as articial stars to draw a distinction between the mere practitioner and the sci-entist. The average practitioner was, as Pritchard put it, not disposed to enter into a scientic scrutiny concerning the causes of their [the microscopes] demerits.38 Only the man of science could be expected to have an interest in the mechanism of


the microscope. The practitioner could well be content with knowing how to use his device and how to check whether it was in proper working order.39 Pritchard thus exploited the reference to test objects as rhetorical device to widen the gap between the mere practitioner of microscopy and the real man of science.

Yet another context for test objects opened up in the mid-nineteenth century as the microscope became established in clinical medicine.40 Notably, clinicians recommend certain body uids and tissues as test objects instead of parts of plants and insects. The Edinburgh physician and microscopist John Bennett, for example, stated in his Introduction to clinical medicine of 1853: if the microscope shows with clearness the epithelia scales, the structure of the salivary globules, their nuclei, and contained molecules, you may be satised that the instrument will exhibit all the facts with which, as medical men, you have to do.41 Bennetts statement also shows that the clinicians goals for testing were slightly different from the goals pursued in the context of microscopical research discussed above. For Goring and Pritchard, one important aspect of testing was to establish the highest powers of individual microscopes. But for Bennett, the guiding question was whether the microscope at hand was suitable in the sense of good enough for routine purposes, namely the search for pathological changes of body uids and tissues and signs of disease.

The increased use of microscopes as research tools as well as the gradual establish-ment of the microscope in clinical routines were complemented by the introduction of practical exercises in microscopy as part of medical education, especially from the 1840s onward. Test objects served as training tools for beginners and students in the classroom.42 Frey recommended test objects as discipline for the beginner since their resolution is by no means easy, and with them the accurate adjustment of the focus, and the skilful application of the illumination, may be learned.43 A footnote in William Carpenters classic microscopy manual The microscope and its revelations also testies to the discussions about how best to employ test objects to acquire the necessary skills. Carpenter noted:

It has been urged that the acquirement of the power of displaying difcult Diatom-tests, is a valuable gymnastic for the training of Microscopists; but the experience of the Author, and of every Biology teacher he knows, is that a much better training for the Student is to begin with the study of such easy objects as afford him the experience which is absolutely essential that he should acquire in the rst instance, and to proceed gradually from these to the more difcult, gaining new knowledge at every stage.44

Working with test objects was not only gymnastic for the neophyte but also a sport for scientic gentlemen and microscope enthusiasts, and as such it became an end in itself. Goring already alluded to this when he compared trying microscopes against each other with horse racing.45 Especially in the second half of the nineteenth century, when diatoms delicate marine algae became common test objects, many diatomaniacs engaged in diatom-hunting to be able to stage even more difcult trials for their microscopes.46 Carpenter commented dryly: It is assuredly neither

124 JUTTA SCHICKORE

the only nor yet the chief work of the Microscope (as some appear to suppose) to resolve the markings on the siliceous valves of Diatoms.47 Carpenters reviewer was even more explicit. He simply declared these amateurs were not microscopic observers. The only thing they proved was that they had a purse in their pockets. But doing research with an inferior microscope testied to brains in ones head.48 Once more, test objects served as a means to draw boundaries between the amateur and the professional. But while in previous years, the application of test objects was the mark of the profession, excessive usage was now the mark of the amateur.

This section has shown that in the second third of the nineteenth century, multi-ple contexts for testing microscopes opened up. Gorings initial denition of test object already brings together two roles of test objects: the comparison of different instruments (which one is showing more and better?) and the quality check for an individual instrument: the assessment of its pretensions, defects, and excellencies. Several additional contexts and aims for testing soon unfolded. In each context, specic test objects came to be used for specic goals. Particularly delicate natural objects became reference points in the competitions among makers and users for technical superiority. In research contexts, certain nely-lined parts of plants and animals served to evaluate the good working order as well as the highest powers of the microscope at hand. In clinical contexts, common uids and tissues, which were representative for microscopic objects in routine clinical checks, showed that the microscope at hand was good enough to perform routine tasks. Such common objects also became training tools for students. Articial objects specically designed for testing the optical system served makers and users as indicators of faults and fur-ther incentive for technological improvement. And as we have seen, an unintended consequence of the introduction of test objects was that they became instrumental in the drawing of boundaries between different groups of microscope users: amateurs, [mere] practitioners, and [real] men of science.

GOOD TEST OBJECTS FOR GOOD MICROSCOPES

One would expect that using tests gradually became a stable and routine practice once the practitioners had gured out which test objects were the best indicators for good microscopes. But while the purposes for testing multiplied, the properties that test objects were intended to evaluate became progressively specic and precarious. Test objects turned into objects of study and quickly became rather recalcitrant little things. While the overall merit and usefulness of test objects were never in doubt, the preoccupation with test objects and testing practices gradually transformed on the features that were considered essential for good microscopes and that were thus required of good test objects. New obstacles and challenges for the practitioners opened up on new levels, and the practice of testing never fully stabilized.

Initially, a diverse set of organic objects were used for the purpose of testing and comparing microscopes, ranging from bats hairs and plant seeds to the feet of ies and the jaws of cheese mites. Soon, several of these objects were dismissed, while others became common currency. Some of the reasons for this were clearly


pragmatic. Certain objects were set aside because they were costly; others because they were hard to procure; yet others because they were so difcult to handle. But which objects continued to be used also depended on the changing notions of good microscopes.

What is a good microscope, and what makes a microscopic object a good test object? Reading Gills report of Amicis visit to London, one can get a sense of the early applications of test objects in the context of the social gatherings of microscope enthusiasts. Gill told his readers how he had

witnessed the performance of his [Amicis] microscopes in comparison with others, brought for the express purpose, and with the same test objects; and the result appears to be, that his reecting microscope is a most capital instrument indeed, under moderately magnifying powers; but his refracting one, which is composed of a compound object glass, formed of two achromatic lenses, each consisting of three lenses; a right angled prism, to change the direction of the line of vision from a vertical to a horizontal one; and an astronomical eye-glass, greatly exceeds it for every purpose, and is indeed a superior instrument.

He reported that the result of the comparison appeared to be in favour of Profes-sor Amicis refracting microscope, and of Tullys [sic] achromatic and Cuthberts miniature-copy of the professors reecting microscopes. Gill pronounced these instruments the most perfect microscopes hitherto constructed.49

Gills report shows that the assembled gentlemen used test objects to compare types of microscopes from particular makers, and that the goal was to establish which type was superior to the others. The subsequent issue of the journal tells us how Gills own single microscope made by Varley came under scrutiny. The report suggests that in the early stages of using test objects, one main goal of testing was to check whether ones microscope could display at all the pattern or surprising features of a test object. Gill reported that he was invited by another gentleman, Mr. Carpenter, to bring his Varleys single microscope to his house , in order to prove its powers by the inspection of a few test objects.50 Gill then described the tests that his microscope had undergone. Some of these are familiar from Gorings publica-tions, others more extravagant: at rst Gill had to try the feathers of the Menelaus, the cabbage-buttery, and others, after which Carpenter presented Gill with a more curious object, a cheese mite, placed upon a slip of glass, and covered with a slice of talc; telling him, that he had endeavoured to discover the anatomy of the mite, but not mentioning in what respect. But Gills microscope stood the challenge:

... upon submitting the object to the microscope, the Editor quickly found, that two sets of teeth and the jaws were most beautifully displayed on each side of the object. On mentioning this, Mr. Carpenter said they were what he wished to try the power of the microscope with, but would not have previously mentioned them, but let the Editor have found it out himself. The teeth are, as may be supposed, exceedingly delicate, but are most distinctly displayed, and form exquisite microscopic objects.51

126 JUTTA SCHICKORE

The case of the cheese mite is signicant because it shows how deeply the practice of comparative trials was rooted in natural history. The microscope offered surprising, unexpected views, and this ability to surprise the viewer testied to the instruments high powers. This kind of test the trial often involved very specic, highly context-dependent knowledge of natural history.

One of the initial goals of testing was simply to resolve the pattern or display the outline or a certain surprising feature of an object, and if one could declare a pattern resolved, the test was considered successful. This was also what Gill and Pritchard had in mind when they sent their test objects to France. But soon, resolving was no longer enough. The quality of the display came under scrutiny, and the practitioners used tests more systematically to establish whether a microscope performed better than another, and to assess the specic features of each instrument.

As Pritchards essay on test objects demonstrates, by the early 1830s, the evalu-ation criteria had already grown more complex. Pritchard sorted test objects into groups according to the degree of difculty with which they could be resolved and according to the specic optical properties they were to evaluate. Pritchards text contains a list of test objects accompanied by a plate with depictions of these objects. The list comprises two kinds of objects, those for testing the penetration and those for testing the denition of an instrument. Pritchard explained that penetration, which depended on the angle of aperture, denoted the instruments capability of showing ne lines, while denition was inversely as the quantity of spherical and chromatic aberration, that is, it indicated the instruments capability of displaying the outlines of the object.52 Generally speaking, lines on the wings and scales of insects tested penetrating power, and the outlines of the hairs of animals and certain mosses ascertained dening power. Each of Pritchards classes comprised easier and more difcult objects, up to the most difcult. Pritchard claimed: When an instrument shows the last class properly, it may be at once pronounced superlative.53 The plate presented the objects correspondingly; they were grouped into two sets according to the optical properties they assessed. Pritchards classication of test objects implies that the most difcult ones those with the most delicate structure were the best, as they singled out the most perfect microscope. Roughly, the term difculty (which was often used in quotation marks) denoted the degree of delicacy of the shape, pattern, and texture of the object, whose display required a microscope with an appropriate degree of perfection. A test object that possessed maximum difculty would tax the microscope to ultimate limits.

But in practice, matters were not so simple. Pritchard also noted that in order to be evaluated effectively, microscopes must be tried on the same object and had to be applied under similar illumination. Otherwise it might well happen that a good microscope may be rejected, and a worse one, by better management, allowed to carry off the palm of victory.54 Embedded in Pritchards statement are three more requirements for successful testing. First, the objects used for testing must be stable55 and uniform. Second, the conditions of observation must be stable, uniform, and appropriate to the task. Third, the observer must have the necessary skills to arrange


the set up and to perform the test. The additional requirements in Pritchards conception of successful testing soon

became matters of concern, and as more and more microscopists engaged with test objects, the real signicance of Pritchards insights became painfully clear. Detmers, for instance, reported that none of the European opticians he had visited had been able to resolve the test objects with their German microscopes and stressed that part of this failure was likely due to bad management of the viewing conditions. He noted he would have been well able to resolve the test objects had he been asked to manipulate the German instruments.

The multiplication of evaluation criteria and the growing importance of compara-tive testing created new challenges to the practitioners because they had to design testing practices that enabled such comparison between microscopes. As Gills report shows, in the early days of testing, microscopists carried their microscopes and test objects with them to the social gatherings so that several microscopes could be tried with the same object. But such direct comparison is not always practical. If the goal was mere resolving of certain features of test objects, it would be enough to send just that object around. But circulating delicate objects and across long distances, too! is not very practical because it takes time, and the objects may be damaged or destroyed on their journey. And if the goal was comparison, it was at least required that an observer travelled with the test objects; and even then, the comparison would be questionable because it hinged on the observers good memory. Chevalier returned to his English friends some drawings of the English test objects as they looked when viewed with his microscope. The drawings served as evidence that the test objects had been resolved. But once the patterns of these test objects were known, it was of course debatable whether the drawings were true to the microscopic images or whether they were simply made to look good. The fact that Detmers went on his trip to Europe equipped not only with test objects but also with photomicrographs suggests that this was a concern amongst practitioners and that drawings were con-sidered doubtful pieces of evidence.

As the practice of testing became more common, instances of types of test objects had to be multiplied in such a way that they remained uniform. Pritchard attempted to secure the uniformity of the test objects by stressing the necessity of a careful selection of those [objects] similar to my drawings.56 With this statement, Pritchard implicitly acknowledged that even if a microscopist chose to use the very same part of an insect or plant as he, Pritchard, did, this would not guarantee the uniformity of the object. No two natural objects are exactly the same, and they do not even resemble one another enough for testing purposes. So if one selected ones test objects so that they resembled Pritchards drawings, one could at least make the objects somewhat more uniform.

But Pritchards way of circumventing the problem of natural variation by providing authoritative drawings cannot have been very satisfactory to the other practitioners. The universal standard of comparison that Pritchards image represented was con-stituted by a combination of the skilfulness of Pritchards drawing, the competence

128 JUTTA SCHICKORE

of the engraver, and the quality of the print. Goring and Pritchards depictions of test objects underwent numerous reprints that did not leave the images unaltered. Besides, how could one nd out whether the test object at hand really looked like Pritchards (after all, the details could only be seen with a microscope), and why indeed should one adopt Pritchards exemplars as authoritative models to copy from?

Circulating designated tests and providing depictions as models for good test objects or successful resolution are different activities directed at a common goal: to eliminate variations in test objects. Yet in the longer term, using natural objects was not considered completely satisfactory. The problem of variation in natural objects was tackled in a novel fashion when articial test plates were introduced. In 1846, some fteen years after the publication of Pritchards essay on test objects, the German engineer Friedrich Nobert found a new way of obtaining uniformity across test objects.57 He devised a ruling engine to make nely scaled glass gratings. This technique could also be used to produce test plates for microscopes. The plates had bands of lines, each with specic, increasingly smaller distances between the lines.58 Nobert explicitly stated that mechanical production of test plates was advantageous because it took care of the problem of variability.59

The invention of test plates has been celebrated as notable technological advance-ment for microscopy.60 In hindsight, it appears that the mechanical production of standard test plates did indeed accelerate the dynamics of microscope technology. Microscopists published lists of microscopes by different makers together with information about the highest band of Noberts test plates that they had been able to resolve.61 John Mayall, a near contemporary of Nobert, noted that the nest gratings that Nobert produced always exceeded the resolving power of available instruments, thus challenging the microscope makers to improve the performance of their products.62

However, the introduction of articial test plates did not bring the debates and activities of securing the uniformity of test objects to a satisfactory conclusion. Several practitioners found that Noberts test plates also have the fault of not being identical.63 This is not surprising given the extremely intricate manner of their pro-duction.64 Contemporaneous sources suggest that Nobert removed from Greifswald to Barth because the vibrations from the urban environment impeded his work. For the same reason, in Barth, Nobert took to working during the night.65 If the grating of test plates was so sensitive that it could be affected by everyday city bustle, it is no wonder that variations of the lines could not be entirely avoided.

Actually, the problem of natural variation was not the only reason why articial test plates were welcomed. The other advantage of the rst mechanical test plates was that the lines were ner than those on most natural objects, which meant that test plates were more difcult. I noted earlier that the rst natural test objects had helped produce optically improved microscopes. The improved instruments required increasingly difcult objects. Mechanically produced plates and natural objects continued to compete for the highest level of difculty. In the late nineteenth century, the German microscopist Heinrich Frey pointed out that certain natural


test objects had been abandoned with the increasing perfection of microscopes. In particular, all those which were recommended before 1840, all the various hairs and scales of butteries and wingless insects, may be regarded as conquered ter-ritory. To attempt to test a rst-class modern microscope with these expedients of a former epoch, would be an insult to the optician from whose establishment the instrument has proceeded.66 Frey illustrated this with the fate of the Janira scale, Amicis favourite test object.67 In the mid-1840s, the botanist Hugo von Mohl had recommended this object as a very difcult test. Frey recalled how he as a student equipped with a Schieck microscope had to vex and trouble himself to obtain only a passable view of these transverse markings. Decades later, by the time he wrote the instruction manual, an instrument that could not show these markings at a magnifying power of 200 was considered bad.68

Throughout the nineteenth century, keeping pace with technological advancements and nding more and more difcult test objects both remained concerns. John May-alls paper on Noberts ruling machine, which was delivered to the London Society of Arts in 1885, suggested that a major motivation for adopting articial test plates had in fact been the scarcity of natural test objects with extremely ne lines.69 In the 1850s, when diatoms were discovered as test objects, they at rst outperformed Noberts test plates. Moreover, Noberts test plates were barely affordable while diatoms were rather cheap. Diatoms and Noberts test plates remained valid options throughout the century because of the very similar degrees of difculty of these objects. Ngeli and Schwendener included in their 1867 manual of microscopy a table displaying the widths of Noberts test bands, adding that the table showed that Noberts test plates were only a little less difcult than the most delicate organic objects.70 Noberts own comparative table of 1861, published posthumously in 1882, shows the opposite.71 In 1872, Heinrich Frey again observed that only the markings on Noberts later test plates were delicate enough that markings as ne as those of the Diatomaceae have been made by art.72

Finding sufciently delicate microscopic objects was not the only problem the microscopists faced. They gradually realized that even the most fundamental require-ment for test objects could not always be met: obviously, the principal requirement for a test object is that its structure be known. If one is not thoroughly familiar with the ner structure of the test object one is applying, the object will not be very useful for establishing the quality of a microscope, because one will never be in a position to judge whether the microscope has met the goal of displaying the object correctly or indeed better than another. Yet in several cases, the true structure of the test objects was very much in doubt. For instance, in 1858, the science instructor Friedrich Rein-icke recorded in his microscopy manual the disputes among several practitioners as to whether the diatom Pleurosigma angulatum showed lines or whether these lines were separate dots.73 In 1870, the microscopist Joseph Bancroft Reade lamented that there were as many descriptions of the diatom Pleurosigma as there were observers.74 The structure of the common test object Lepisma saccharina (the silversh) was also much contested. In 1873, the American naturalist G. W. Morehouse pointed out that

130 JUTTA SCHICKORE

even though its scales had been for many years subjected to most careful and critical examination by the most competent observers and the best microscopes, the true character of its markings still remains a disputed question. The disagreement arose, he said, partly from the complex nature of the markings themselves, and partly from the different conditions under which they have been seen.75

Articial test objects did not fare better. One should not assume that because their lines were man-made and engineered, they were easier to understand. On the contrary, as a note in the 1869 Popular science review indicates, there were compet-ing methods of counting the lines on Noberts test plates, and the descriptions of the methods suggest that the task was rather precarious.76 Mayalls talk on Nobert and his work points out that the capability of resolving the highest bands hinged on the method of illumination.77 In the 1870s, microscopists still recorded diversity of opinions about the resolution of the latest test plates.78

Despite all these obstacles, the microscopists never abandoned the practice of using tests. On the contrary, they took the sorry state of things as an occasion to investigate closely the various conditions of viewing test objects and the causes of the discrepancies in the descriptions.79 They established, among other things, that the way of preparing and mounting (dry v. wet) had an inuence on the ease with which the markings on the test objects were seen.80 They also examined how the perform-ance of the human eye constituted the practice of testing. It had long been noted that microscopic vision was different from vision with the unaided eye and that this might be a cause of error. Similarly, the unusual viewing situation might impede the study of test objects.81 The American microscopist Lucian Howe, for example, drew atten-tion to a common imperfection of the eye, astigmatism. Because of the eyes in-built imperfection, ne parallel lines could not be fair tests for microscopes. Referring to the great authority F. C. Donders, the Dutch physiologist, Howe stressed that astigmatism was a feature of normal eyes. If two people looked at Noberts test plates or nely-striped diatoms, one of them might thus see the lines perfectly well while the other was incapable of seeing them. This person would of course pronounce the microscope inferior. Howe claimed to have encountered frequently such a difference of opinion. In these cases, it was imperative that the object on the microscope stage should be turned through an arc of 180 degrees. By revolving the object in such a way that the horizontal lines became vertical and vice versa, they passed through every position in which an observer could possibly see them.82

Eventually, the microscope even became a tool to evaluate the limits of the testing procedures for microscopy. In their manual of microscopy, the botanists Carl Ngeli and Simon Schwendener noted that the visibility of the dark lines of test objects depended on the eyes acuity. Therefore, they claimed, the purported testing of the microscope is essentially a testing of the eye.83 Ngeli and Schwendener based their study of test objects on micrometric measurements of the size of light-sensitive elements of the retina.

In his study of late Victorian microscopy, Gooday points out that the illumination was generally considered one of the most important and most precarious parameters


for microscopic observation. It is thus not surprising to nd that in the context of testing, the practitioners also paid particular attention to illumination the soul of the complex body, with all its ingenious mechanism, appliances, and powers.84 In his article on test objects, Reade bemoaned that while the practitioners had numerous kinds of illumination apparatuses to choose from, none of these was really satisfac-tory. He then set out to show that his method could bring out the real structure of various diatoms.85 The German microscopist Frey too embarked on an investigation of the effects of illumination. His work even showed that the difculty of one and the same object, the diatom Pleurosigma angulatum, varied in different illuminations and different kinds of objectives: with oblique light, this diatom was suitable to test medium and greater powers, but with good immersion lenses it was important to use central illumination. If immersion objectives were used with oblique illumination, the test became too easy.86

All these examples illustrate the intricate relations between the perceived dif-culty of a test object, technological advancement, and the management of the experimental setting. Morehouses paper on the silversh Lepisma saccharina gives another glimpse of this complex of concerns. In Morehouses opinion, the continu-ing disputes about the true structure of the scales were in part due to the fact that the object itself had coarser and ner markings, a fact that had only become apparent with the technological improvements of microscopes. Initially the coarse ribs of the scales of the insect had appeared easy to resolve. However, as Morehouse showed in his paper, improved microscopes and better management of the illumination had revealed delicate structures severely taxing the powers of the nest objectives in existence. Morehouse described a whole series of structures in increasing order of difculty, ranging from heavy longitudinal ridges to faint irregular veins.87 But he also pointed out that some of the markings that had been described by others were due to optical illusions produced by inappropriate illumination. So it was no wonder that observers using different instruments with different illumination had come to contradictory conclusions about the structure of the test object and its degree of difculty.

Investigations of the effects of light continued well into the twentieth century, as an essay from the 1909 issue of the American natural history journal Midland naturalist demonstrates. The essay begins by conrming the overall importance of testing: Even the amateur microscopist has come to recognize that the value of a microscopical objective depends on the ease with which it will resolve with perfect clearness certain test objects usually diatoms with delicate markings. But the focus of the essay is on the problems of displaying the markings of a particular diatom commonly used as tests (Amphipleura pellucida). The author announced that in lengthy experiments with various objectives he had accidentally found that the illumination was critical for the correct display of the delicate lines of the object. He obtained the best results when the incidence of the direct sunlight took place as nearly as possible parallel to the stage, the microscope being inclined for the purpose, and at right angles to the striae.88

132 JUTTA SCHICKORE

The manifold investigations of test objects, of their difculty, and of the condi-tions of testing had another important consequence. The very concept of difculty became questionable. Not only was it hard to nd suitably delicate or difcult microscopic objects, but difculty as such turned out to be a problematic notion. Only at rst, the term difculty, or rather degree of difculty, was used simply as an indicator of the degree of delicacy of the object, and, by implication, the power of the microscope that could display it. Difculty soon became a puzzle: what exactly did it mean that a test object was difcult, what was the power required to display it, and was the ability to display the most difcult objects really the most desirable property in a microscope?

Already in the 1830s, Pritchards essay showed that test objects could be difcult in different respects. Pritchards ordered list of test objects demonstrates that by that time, test objects already had more specic functions than Gorings and Gills delicate specimen had had. According to Pritchard, test objects assessed particular optical properties rather than the overall power or degree of perfection of an instrument: some test objects rated penetrating power, others dening power.

As I noted above, Pritchard drew attention to the fact that these properties were independent from one another in such a way that the instrument might possess a very considerable approximation to perfection in the one, and fall short in the other, or vice versa, or might be perfect in both.89 Each microscope manifested the two optical properties to different degrees. Only gradually did the microscopists come to appreciate the implications of the insight that the difculty of test objects was in fact an indicator of different properties. But this insight gave rise to another question: Which of these properties was the most desirable? Pritchard had assumed that the most perfect microscope was the one that combined the highest degree of denition with the highest degree of penetration. However, it became clear that extremely high penetrating power was usually obtained at the expense of dening power and vice versa. So one had to determine which of the two properties was the more important, and how much spherical and chromatic aberration was still accept-able in a microscope.

The situation became even more complicated when it turned out that the multi-faceted capacity to display difcult objects was not the only desirable power of a microscope, and indeed that the capacity to display the most difcult objects might not be the most desirable property for a working microscope. As early as 1829, Jacquin cautioned that while the quality of a microscope could only be determined by test objects, those makers who designed their instruments exclusively to display these objects in the best possible way were mistaken. After all, the microscopist would not only want to test but to use his microscope.90 And the Austrian microscopist Joseph Johann Pohl opened his essay on test objects: The test of a microscope with so-called test objects decides only about the optical value, while other important properties requisite for the practical use of the instrument are disregarded.91 Other practitioners also pointed out that test plates measured selected features of the instru-ment and only in the particular part of the aperture that was being utilized. Mayall


implied that it was a mistake to regard the resolution of a test band with very ne lines as a sufcient test of the quality of an objective because this fact merely dem-onstrated that one zone of the objective was free from aberration. If the rest was not, this objective was in fact of inferior quality.92 Therefore test plates might not give an altogether fair assessment of microscopes; they were certainly not the ultimate measure of good microscopes.

Carpenter then distinguished ve desirable attributes of a good object glass for microscopes: (1) its working-distance, or actual interval between its front lens and the objects on which it is focused; (2) its dening power, or power of giving a clear and distinct image of all well-marked features of an object, especially of its boundaries; (3) its penetrating power, or focal depth, by which the observer is enabled to look into the structure of objects; (4) its resolving power, by which it enables closely-approximated markings to be distinguished; and (5) the atness of the eld which it gives.93 Carpenters approach shows that by the late nineteenth century, the notion of a good microscope had become quite intricate. Moreover, both the question of what a good microscope was and the question of how to assess whether particular instruments were good were wide open. Carpenter claimed that test objects could assess only two of the properties he considered crucial: denition (the power of giving a clear and distinct image of the markings of an object) and resolution (the power of distinguishing closely approximated markings); and anyway, most tests were only for resolution, that is, for angular aperture. Other microscopists took a different approach. The American J. J. Woodward, for instance, favoured versatile multi-purpose test objects such as frustules (cell walls of diatoms), which he thought could test several properties at once. They served as a valuable unit of comparison between different objectives, the distinctness with which the striae are shown indicating the denition of the glass, the manner in which the edges of the frustule are seen while the mid-rib and striae are in focus showing the degree of penetration, and the appearance of the ends of the frustule when the centre is in focus giving a fair idea of the atness of the eld.94 These examples show that in the late nineteenth century, the whole procedure of testing had become extraordinarily complex. Not only was the practitioner required to pit dening power against penetrating power but there was in fact a whole array of features that needed to be weighed, and how should one go about it? What were the most desirable features of a microscope, and what test objects were best suited to assess them? Opinions were divided. While the practitioners were only too aware of this complexity, no overall consensus had emerged about good microscopes, good test objects, and ideal viewing conditions.

THE CAREER OF TEST OBJECTS IN THE NINETEENTH CENTURY

How can one characterize in general terms the career of test objects in the nineteenth century? In many ways, it was a success story. Initially a number of delicate objects served as rough-and-ready comparative measures for the optical quality of instru-ments. As the testing contexts multiplied, these objects were sorted into different kinds for different purposes. Test objects were, and continued to be, relevant in the

134 JUTTA SCHICKORE

more informal exchanges between microscope enthusiasts and instrument makers as tools to compare microscopes and to evaluate their highest powers. These interactions inspired technological advancements, which allowed for the use of more sophisti-cated test objects; and that, in turn, encouraged further technological progress. The reference to test objects became part of the researchers arguments for the validity of their results. In the latter half of the nineteenth century, the test objects had become established tools in the clinic and in educational contexts.

However, the career of test objects cannot easily be described in terms of stabili-zation and establishment of routine procedures. Here my conclusions diverge from Goodays. Gooday writes that in the longer term, late Victorian microscopists (and physicists) were able to enhance these conditions [of their research] by reengineer-ing their working environments to achieve greater phenomenological orderliness: physicists by acquiring specially purpose-built laboratories, and microscopists by installing auxiliary devices such as the achromatic substage condenser into the very structure of their instruments.95 My study of the history of test objects and testing procedures suggests that the test objects and the viewing conditions were never fully stabilized, and testing never became a uniformly routine practice.

The article in the Midland naturalist of 1909 that I mentioned earlier is an indicator of the general attitude toward test objects at that time. It shows that at the beginning of the twentieth century, the general usefulness of test objects was beyond doubt. However, the true structure of these objects, the right viewing conditions, and the concrete goals for testing were still matters of concern and investigation. Until well into the twentieth century, practitioners debated what certain test objects really looked like. Even then, the activities and investigations of the testing practices and conditions had not fully stabilized the testing procedures. And if one measures the practice of testing with the requirements the practitioners expounded, their practice constantly fell short of their ideals. For instance, the uniformity of test objects was never fully guaranteed. It was obvious that natural objects are never completely uni-form, and it was also clear that the procedures to engineer articial test plates were never totally exact and reliable.

The continuing attempts to make test objects and the testing situation more uni-form, to establish the true structure of the test objects, and to identify objects of an appropriate level of difculty even had adverse effects. The more the practitioners engaged with difcult test objects, the more diverse descriptions of test objects were offered. The more insights were gained in the viewing conditions for test objects (and microscopic objects more generally), the more idiosyncratic the testing procedures became, as individual practitioners sought to design as well as they could stable test-ing situations for themselves. There were signicant tensions in the microscopists deliberations about test objects, because the practitioners declared that standard, uniform testing procedures were important while proposing and rening their own personal preferences and practices.

From early on, practitioners demanded that test objects had to be universal in order for test results to be comparable. Jacquin, for instance, lamented that each microscopist


had favourite objects depending on his specialty: the physiologist used blood globules, the botanist pollen, and so forth. But it was desirable that one came in general to an agreement about certain objects that are not only suitable because of their structure and durability but also uniform and obtainable everywhere.96 Goring also insisted that the tests be performed by a neutral judge, the rst rule of testing being that no mans judgment is worth a straw, relative to his own instrument, even in a case where there is no money to be lost or won by it.97 On the other hand, authors of microscopy books and articles demanded that the practitioners familiarize themselves with a few test objects so that they were always in a position to judge the qualities of the micro-scopes they encountered.98 Indeed, many practitioners seemed to have their favourite objects, which they used in conditions they found most favourable.99 In 1861, the Austrian microscopist Pohl explicitly regretted that no denitive classication of test objects according to common principles had been agreed upon. His survey lists the different classes of test objects recommended by Jacquin, Goring, Chevalier, Mohl, Schacht, Grifth, Robin, Carpenter, Bailey, and Amici. Pohl then argued for his own classication, which sorted test objects into three kinds: those testing for penetrating power, those testing for dening power, and those testing for chromatic and spherical aberration.100 In a similar vein, Jabez Hogg noted that there were almost as many methods employed for bringing out the markings on test-objects, as there are skilled and efcient microscopists, each one preferring his own, simply because he has been at great pains in working it out with the utmost nicety.101 Hogg also illustrated his complaint with descriptions of several different methods of displaying test objects as they had been proposed by various microscopists.102 His survey highlights the problems that had to be overcome and the specic problems each test object posed: not only did each observer apply individual methods but the methods even differed for each test object. One might therefore say that if at all, testing procedures were stabilized at the expense of generality.

If one compares late nineteenth-century approaches to testing microscopes with the early accounts of test objects and their usages, several signicant differences emerge. For Gill and Pritchard, it was clear that the most desirable microscope was the one that displayed the most difcult test objects to perfection. But the initial require-ment for good test objects that was suggested by the comparison with astronomy namely, that the test objects should have delicate patterns was merely a pre-liminary step in the identication of suitable test objects. Discussions continued about what exactly the right requirements were for good test objects in different contexts, how they could be met, and if they could be met at all. Ultimately, these discussions changed both the conceptions of good microscopes and the requirements for good test objects. In the late nineteenth century, the practitioners stressed that what was the most desirable property of a microscope depended on the purposes for which the microscope was used.

My survey also indicates an explicit acknowledgement of the limitations of tests and testing. Many practitioners doubted that highest power was generally the most desirable feature of a microscope and by implication that highest difculty was

136 JUTTA SCHICKORE

really the greatest merit in a test object. Carpenter put it quite bluntly: As in regard to the qualities of Objectives, so in respect to Illumination, may it be condently asserted that the solution of the most difcult Biological problems to which the Microscope has been yet applied, has been attained by arrangements by no means the most favourable to the discernment of the markings on Diatom-valves or the lines on Noberts test plate; and that, conversely, the arrangements specially effective for the resolution of the most difcult lined test have not, as yet, been shown to have much value in Biological investigation.103 In other words, it was a complete mistake to assume that an object which will show certain Test-objects, must be very superior for everything else to a glass which will not resolve these the quali-ties which enable it to resolve some of the more difcult tests, not being by any means identical with those which make it most useful in all the ordinary purposes of Scientic investigation.104

Like Pritchard, Carpenter presented test objects for low, medium, and high powers: among them injected preparations of a frogs lung and Morpho menelaus (for low powers), certain diatoms and hairs for medium powers, and test plates and more difcult diatoms for high powers. But Carpenter singled out an object from the medium range as particularly useful to evaluate the working power of the microscope. He recommended the Podura scale, stressing that a lens that brings out its mark-ings satisfactorily will suit the requirements of the ordinary working Microscopist, although it may not resolve difcult Diatoms.105

Many microscopists around 1900 explicitly stated that for most tasks extremely high magnication was not necessary and indeed detrimental. These practitioners cautioned that we know, indeed, with every increase in this direction, how liable we are to encounter unforeseen errors and exaggerations.106 In everyday practice, it was therefore unnecessary and even counterproductive to utilize the most difcult test objects. In most cases, moderate magnication was sufcient; and test objects simply had to show that microscopes worked well and showed objects clearly and distinctly in this middle range.

In addition, it turned out that test objects could be difcult in different ways, depending on the specic optical property that was required for them to be displayed. Because certain features of a microscope could only be improved at the expense of others, there was always a trade-off of desirable properties, and it was a contingent question what the most desirable properties of the microscope were. In any case, it became clear that a microscope that was superlative on all counts would never be available. Working microscopes would always have certain limitations. Test objects and testing were intrinsically limited, too. In the chapter on test objects and micro-scope testing in his 1877 microscopy manual Julius Vogel pointed out that because the appearance of these objects depended on minor details [Nebendinge] such as oblique rays and revolving microscope stages,107 test objects were not an absolute measure of the quality of microscopes.108

Only in the context of those research projects that pushed the boundaries of instrument-aided vision was it imperative to secure the highest powers of microscopes


with the most difcult tests. And in these situations lurked the most formidable problem of microscopy: when microscopists ventured out into new domains of the smallest things, the highest powers of microscopes had to be secured with the most difcult objects. But these test objects were microscopic objects at the limit (and indeed beyond the limit) of the optical capacities of the very instrument whose performance they tested. Here, the microscopists encountered a second kind of trade-off, a trade-off between epistemic gain and risk. The trade-off at the limits of microscope-aided vision could not be circumvented because those test objects that had become familiar and routine objects were no longer suitable to explore the microscopes ultimate powers.

Nineteenth-century microscopists researching at the limits of instrument-aided vision assessed their research tools with standards that were not fully known to them. Not until new types of microscopes became available whose powers considerably exceeded the light microscope (notably the electron microscope) could the uncertainty about the most difcult test objects for light microscopes be removed and their application vindicated. But precisely because new technologies had become available this vindication was obsolete. In 1947, for instance, the author of an article on diatoms motivated his investigation with the remark that the diatom Amphipleura pellucida had frustrated the efforts of some of the best microscopists for the last seventy-ve years.109 He had thus set out to clarify the structure of that diatom with the help of the electron microscope. So in a sense, the best validation of light microscopes could be obtained only when it was no longer needed. But as more recent articles on test objects demonstrate, the very trade-off of epistemic gain and risk repeated itself for the electron microscope.110

CONCLUSION: LONG-TERM HISTORY OF EVALUATION PROCEDURES OF MICROSCOPES

It remains to reect on the role test objects played in the long-term history of evalua-tions of microscopes. Needless to say, within the connes of this article, I cannot trace in detail the history of evaluation procedures and requirements for good microscopes. In what follows, I merely highlight some signicant features of evaluation criteria and practices in the early modern period and the late eighteenth century and contrast these with the principles and practice of using test objects in the nineteenth century.

Debates about the merits of the microscope are as old as the microscope. But the nature of these concerns changed signicantly. Robert Hookes Micrographia of 1665, for example, introduces microscopes and telescopes as powerful extensions of our senses. Hooke argued that optical instruments could serve as tools to remove the imperfections of man and correct the falsehoods of scholastic philosophers who relied solely on the strength of reason.

Hookes reections appear in the preface of his book, not in connection with the concrete ndings described in the main text. Most of the reections concern general metaphysical and epistemological issues. Hookes comments about the values and virtues of instruments can be read and have been read as attempts at present-ing a compelling image of the New Philosophy that was then promulgated by the

138 JUTTA SCHICKORE

Royal Society and as reections of Hookes devotion to Christianity.111 They tell us of Hookes hope that the articial extension of the senses might restore mans uncor-rupted view of nature. Hooke also offered comparative assessments of microscopes. But he compared kinds of microscopes rather than individual instruments. He compared the single with the compound microscope and recorded his trials with other kinds of microscopes tted with one piece of Gla, both whose surfaces were plains or with plano-concave lenses, and with instruments made of Waters, Gums, Resins, Salts, Arsenick, Oyls, and with divers other mixtures of watery and oyly Liquors, all of which he found to be inferior to his compound microscope. The text does not tell us anything about how he had compared the microscopes and what the principles of comparison had been. Hooke merely announced that he had found none generally more useful than his compound microscope. Hooke conceded that his microscope showed objects dark and indistinct. This inconvenience was, he said, inseparable from Spherical Glasses. He recommended that strong light be used to improve the performance of the instrument. The text indicates that he believed in the possibility of improving optical instruments to perfection. Referring to Descartes, Hooke noted that much greater perfection of Opticks could be expected from elliptical glasses.112

The practice of comparing microscopes to one another in order to nd the best had become common in the eighteenth century. Several writers on microscopy reported the results of these comparisons. For instance, the naturalist D. Pelissons 1775 article in the journal of the Berliner Gesellschaft Naturforschender Freunde [Berlin Society of Scientic Friends] compared an English, a Dutch, and three German microscopes with respect to focal length and magnication.113 Notably, Pelisson had not made all the assessments himself. He regretted that he had not had occasion to try the English instrument, so he related the information he had received from a friend who owned one. The article includes a table with data obtained from different microscopes (the composition of the lenses, their distance from one another, and the focal length). However, the table is not connected to the assessment of the performance of the microscopes. The comparative assessments in the text are qualitative. It is also strik-ing that more than one-half of the article describes observations of infusorians, and these reports are completely disconnected from the comparative assessment in the earlier part. Pelisson no longer dwelt on the merit of the microscope as prosthetic device. But like Hooke, he compared types of instruments. Again, it is not quite clear how the comparison was made and it was, in part, based on hearsay. As in the Micrographia, there is no tight connection between the assessment of the instruments and the observations reported in the remaining part of the text.

In his study of eighteenth-century testing procedures, Marc Ratcliff identies different contexts of testing, in particular, advertising and so-called admission. For the purpose of advertising that is, promoting his microscopes, a maker invited other practitioners to compare their microscopes to his. Ratcliffs study thus shows that the social contexts and aims of trying microscopes against one another were very similar in the late eighteenth and nineteenth century. In the eighteenth century, testing was an integral part of the interactions between makers and users


and a motor of technical advancement. Ratcliff also shows that national rivalry was a driving force of instrument testing.

Investigators who obtained a new microscope routinely carried out admission tests; that is, they compared its performance with the other microscopes in their posses-sion.114 These comparisons were carried out privately and with objects that happened to be at hand and of interest to the researcher. Ratcliff argues that one purpose of this test was to prepare the ground for an epistemic triangulation procedure: practitioners often presented different kinds of instruments, which were used on the same object in order to piece together the true image of that object.115

As Gills report of Amicis visit to London demonstrates, the early applications of test objects were a continuation of late eighteenth-century evaluation principles and practices. However, in the course of the nineteenth century, the introduction of test objects changed the principles and practice of testing in at least three ways. First, test objects helped shift the attention from concerns about the merits of the microscope as such and the advantages and disadvantages of types of microscopes or microscopes from certain makers to concerns about the capability of the concrete instrument at hand. The name of the maker or the type of microscope no longer guaranteed its value. The individual users were required to evaluate the performance of the tools they were applying, and the test objects provided the means to do this.

Secondly, the practice of testing with designated test objects as well as the con-tinued attempts to understand and improve the testing situation had an impact on the idea of a good microscope and good test objects. In the course of the nineteenth century, the microscope turned from an instrument with a certain degree of perfec-tion to a complex device with specic pretensions, defects, and excellencies. The overall quality of a microscope was always the result of a trade-off of different optical properties. The evaluation of microscopes involved intricate assessments and a variety of testing tools, which also possessed a range of properties, and which had to be marshalled as required by the purposes the microscopes should serve.

Thirdly, microscopists ultimately had to content themselves with imperfect instruments and evaluative procedures that were not completely satisfactory. My overall conclusion is thus different from Goodays assessment of late Victorian microscopy. While he emphasizes the gradual stabilization of the working condi-tions, I wish to highlight that the practitioners encountered limits of stabilization as they conducted their research at the boundaries of instrument-aided vision with instruments and testing procedures that were not fully understood. Test objects prompted the notion that microscopes were intrinsically limited and would forever remain imperfect.116 Moreover, microscopic investigations would always include a trade-off between epistemic gain and risk. This situation changed in the twen-tieth century, when new kinds of microscopes became available that allowed for multiple determinations of observations whereby results obtained with different types of microscopes could be comparatively assessed. But at the limits of these technologies, researchers again faced the very same trade-off.

140 JUTTA SCHICKORE

ACKNOWLEDGEMENTS

Portions of this material were presented at the Berkeley-UCSF Colloquium, UC Ber-keley; the Department of History and Sociology, University of Pennsylvania; and in the history of science reading group at the Institute for Advanced Study, Princeton. I am grateful to the members of the audiences as well as to Domenico Bertoloni Meli, Elihu Gerson, Heinrich von Staden, and two anonymous referees for History of sci-ence for their helpful comments on drafts of this article. I completed this essay while I was a member of the Institute for Advanced Study, Princeton. Generous funding from the Mellon Foundation is gratefully acknowledged.

REFERENCES

1. This article draws on material from my book The microscope and the eye: A history of reections, 17401870 (Chicago, 2007), chap. 5. In the present essay, I develop those aspects of the story that go beyond the books scope as well as the general implications of the case.

2. See G. Gooday, Nature in the laboratory: Domestication and discipline with the microscope in Victorian life science, The British journal for the history of science, xxiv (1991), 30741, and Instrumentation and interpretation: Managing and representing the working environment of Victorian experimental science, in B. Lightman (ed.), Victorian science in context (Chicago 1997), 40937.

3. Gooday, Instrumentation (ref. 2), 423.

4. My discussion ranges across different countries, and for the most part, I do not take national differences into account. This would be a topic for further, more in-depth studies of the specic activities surrounding test objects. My aim in this paper is to capture the overall dynamics of the practice of using test objects.

5. W. Herschel, On the power of penetrating into space by telescopes; with a comparative determination of the extent of that power in natural vision, and in telescopes of various sizes and constructions; illustrated by select observations, Philosophical transactions of the Royal Society of London, xc (1800), 4985. The Oxford English dictionary identies Herschels paper as the original source of the term. Astronomy remained the reference point for writers on test objects, see F. A. Nobert, Ueber die Prfung und Vollkommenheit unserer jetzigen Mikroskope, Annalen der Physik und Chemie, lxvii (1846), 17385, p. 174; H. von Mohl, Mikrographie, oder Anleitung zur Kenntniss und zum Gebrauche des Mikroskops (Tbingen, 1846), 180.

6. C. R. Goring, On achromatic microscopes with a description of certain objects for trying their denition, The quarterly journal of science, literature, and art, xxiii, n.s. i (1827), 41034, p. 418.

7. Goring, On achromatic microscopes (ref. 6), 41819.

8. A. Pritchard, Test objects, London and Edinburgh philosophical magazine and journal of science, ii (1833), 33544, p. 335. This is the slightly abridged version of the chapter on test objects in Pritchards The microscopic cabinet of select animated objects; with a description of the jewel and doublet microscope, test objects, &c (facsimile edn, Lincolnwood, 1987).

9. P. Harting, Das Mikroskop. Theorie, Gebrauch, Geschichte und gegenwrtiger Zustand desselben. Theorie und allgemeine Beschreibung des Mikroskopes, i, 2nd edn (Braunschweig, 1866), 275n. See also H. A. Smith, Memoir of Charles A. Spencer, Proceedings of the American Society of Microscopists, vi (1882), 4974, p. 66.

10. W. B. Carpenter, The microscope and its revelations, 6th edn (New York, 1883), 161.


11. C. Ngeli and S. Schwendener, Mikroskop: Theorie und Anwendung desselben, i (Leipzig 1867), 79, 83.

12. Ngeli and Schwendener, Mikroskop (ref. 11), i, 120.

13. C. M. Vorce, Penetration in objectives: Is it a defect or an advantage?, Proceedings of the American Society of Microscopists, ii (1880), 7075.

14. See, in particular, Jacob Coxs account of the aperture controversy between the American maker Robert Tolles and Francis Wenham on the other side of the Atlantic: J. D. Cox, Annual address of the president: Robert H. Tolles and the angular aperture question, Proceedings of the American Society of Microscopists, vi (1884), 539.

15. T. Gill, On the microscope, Gills technological repository; or, discoveries and improvements in the useful arts, ii (1828), 9599, 13849, 193201, 25764, 32130, p. 96.

16. J. v. Jacquin,Bemerkungen ber Mikroskope und ihren Gebrauch fr Naturforscher, Zeitschrift fr Physik und Mathematik, v (1829), 12960, p. 137.

17. T. Gill, On the microscope, Gills technological repository; or, discoveries and improvements in the useful arts, iii (1829), 114, 7582, 12949, 193206, 25763, 32127, p. 138; Pritchard, Microscopic cabinet (ref. 8), 151.

18. See, for instance, the catalogues issued around 1830 by the Viennese maker Ploessl (G. S. Ploessl, Verzeichniss der gangbarsten optischen Apparate, welche von G. S. Plssl, privilegirtem Optiker in Wien, neue Wieden, Salvatorgasse Nro 321, fr beigesetze Preise verfertiget werden, Zeitschrift fr Physik und Mathematik, iv (1828), 1218; Neues Verzeichniss der gangbarsten optischen Apparate, welche von G. S. Plssl, Optiker und Mechaniker in Wien, neue Wieden, Salvatorgasse Nro 321, fr beigesetze Preise in Conventions-Mnze oder Augsb. Courant verfertiget werden, Zeitschrift fr Physik und Mathematik, vii (1830), 11928), and, in 1840, by the American maker Spencer (cf. Smith, Memoir (ref. 9), 54). See also J. Vogel, Anleitung zum Gebrauch des Mikroskopes zur zoochemischen Analyse und zur mikroskopisch-chemischen Untersuchung berhaupt (Leipzig 1841), 11920.

19. T. Gill, On Professor Amicis and other microscopes, Gills technological repository; or, discoveries and improvements in the useful arts, i (1827), 1619, p. 16.

20. Gill, Amicis microscopes (ref. 19), 18.

21. Pritchard, Test objects (ref. 8), 335.

22. See, e.g., Anon., Review: the microscope and its revelations, London quarterly review, lxvii (1857), 289315, pp. 2901; see also G. W. Royston-Pigott, On a searcher for aplanatic images applied to microscopes, and its effects in increasing power and improving denition, Philosophical transactions of the Royal Society of London, clx (1870), 591603, p. 591; Smith, Memoir (ref. 9), 63; Carpenter, Microscope (ref. 10), 170; W. H. Walmsley, Some new points in photo-micrography and photo-micrographic cameras, Transactions of the American Microscopical Society, xvii (1896), 3409, p. 340.

23. J. S. Bowerbank, Reminiscences of the early times of the achromatic microscope, Monthly microscopical journal, iii (1870), 2815, p. 281.

24. For the competition between English and Continental craftsmen, see M. W. Jackson, Spectrum of belief: Joseph von Fraunhofer and the craft of precision optics (Cambridge, 2000).

25. Gill, On the microscope (ref. 17), 138.

26. Pritchard, Microscopic cabinet (ref. 8), 151.

27. In his 1882 memoir of the instrument maker Charles Spencer, Smith quoted extensively from letters written in the 1840s (both personal correspondence and letters to journal editors). One microscopist found Spencers lenses far superior to those of Chevalier, stating that with them he had been able, without difculty, to see the cross-lines on the Navicula hippocampus (the most difcult test object now known to me, and which was sent to me from England as the test

142 JUTTA SCHICKORE

object par excellence) (Smith, Memoir (ref. 9), 57).

28. Smith, Memoir (ref. 9), 62.

29. The bulk of Friedrich Reinickes booklet Contributions to recent microscopy, for instance, is devoted to a comparison of English and German microscopes, carried out with test objects (F. Reinicke, Beitrge zur neuern Mikroskopie. I. Die Leistungen der neueren Mikroskope und die Prfung derselben. II. Die Leistungen der englischen Mikroskope, gegenber den deutschen. III. Das Einsammeln und Prpariren der Bacillarien. Mit 9 Abbildungen von Pleurosigma Angulatum als Probeoobject (Dresden 1858), chap. 2).

30. H. J. Detmers, American and European microscopes, Proceedings of the American Society of Microscopists, x (1888), 14954, p. 150.

31. Detmers, American and European microscopes (ref. 30), 152.

32. Goring, On achromatic microscopes (ref. 6), 4223.



35. Articial stars are little globules of mercury. See Harting, Mikroskop (ref. 9), 281, for a description of how to make them.

36. Harting, Mikroskop (ref. 9), 283.

37. H. Frey, The microscope and microscopical technology (New York 1872), 56.


39. For example, in their manual of microscopy, the botanists Carl Ngeli and Simon Schwendener pointed out that test objects were perfect practical tools to evaluate the performance of a microscope if one wishes to avoid testing all the peculiarities and errors of construction on which it [the quality] depends (Ngeli and Schwendener, Mikroskop (ref. 11), 119).

40. On the establishment of medical microscopy, see L. S. Jacyna, John Hughes Bennett and the origins of medical microscopy in Edinburgh: Lilliputian wonders, Proceedings of the Royal College of Physicians of Edinburgh, xxvii (1997), 1221; and A host of experienced microscopists: The establishment of histology in nineteenth-century Edinburgh, Bulletin of the history of medicine, lvii (2001), 22553.

41. J. H. Bennett, An introduction to clinical medicine: Six lectures on the method of examining patients; percussion; auscultation; the use of the microscope; and the diagnosis of skin diseases, 2nd edn (Edinburgh, 1853), 66. Bennett referred here to histological molecules, particles that are smaller than cells but not as small as the molecules of the chemists (on Bennetts histological molecules, see J. Strick, Sparks of life: Darwinism and the Victorian debates over spontaneous generation (Cambridge, 2000), 4247).

42. J. E. Purkinje, Mikroskop, in R. Wagner (