9
The Most Confused Knot in the Doctrine of Reproduction

By James Larson

The systems and methods discussed by Gunnar Broberg and John Lesch in chapters 2 and 3 are among the most characteristic and enduring achievements of the Enlightenment. All of these compendia, from the most circumspect regional flora to the most ambitious inventory of the terraqueous globe, were inspired by the belief that nature's own plan is not only fixed, but intelligible, and known in outline. The physical splendor of these quarto and folio volumes proclaims the scientific and social importance of that belief.

This enormous production, focused entirely upon the stable, observable aspects of natural diversity, has always overshadowed another, smaller but equally important literature written by and for specialists. The transactions and memoirs of academies and universities of the second half of the 18th century are full of studies of degeneration, race, hybridization, metamorphosis, and monstrosity—the irrational back side of the picture devoted to stability and order. Naturalists recognized the existence of these other kinds of diversity with reluctance, and tried to limit the effect of their recognition upon the systems of fixed, natural forms. As knowledge about any one kind of divergence from established order accumulated, specialists segregated the subject and treated it as a separate set of problems. These separate studies did not constitute an integrated body of knowledge, not even at the end of the century, but they did move research in a new direction and they were inspired by a common belief. Naturalists were forced to confront the processes constitutive of living forms and to invent new methods for the study of these processes. If some kinds of diversity seemed to threaten the basic units of natural order with dissolution, it was still possible to believe

that they could be made intelligible, and might even throw light upon the world of stable, established order.

This chapter analyzes one of these specialized studies, the experimental series undertaken by Joseph Gottlieb Koelreuter to discover the reproductive limits of plant species. Koelreuter cannot be seen as a typical or representative figure. The man was unknown outside a small circle of savants in Germany and St. Petersburg. The number of persons who understood his work could be counted on the fingers of one hand. His influence on contemporaries was negligible. In both his professional life and his research, Joseph Gottlieb Koelreuter was, in the exact meaning of the word, unique.

No other experimental sequence in natural history of the late 18th century equaled Koelreuter's in extent. According to Robert Olby, he carried out "more than 500 different hybridizations involving 138 species, and examined the shape, colour, and size of pollen grains from over 1,000 different plant species."^[1] It is not just the number of experiments that is unusual: Koelreuter's use of counting and measurement in the analysis of data is unusual. His contemporaries were masters of impersonal, objective observation, and it would be possible to put together an anthology of their work on generation equal in precision to anything ever written. But throughout the 18th century naturalists continued to ignore or resist the use of mathematical methods and symbols in their studies of morphogenesis. The theory of generation, as they understood it, was observational, qualitative, and finalist. The central conceptual tradition was still Aristotelian, and Aristotle's depreciation of mathematics as irreconcilable with the study of final causes was so thoroughly a part of the science that naturalists did not need to justify their resistance to measuring and counting. Almost everyone considered the constitution of living forms so complicated and subject to so many conditions that their complexity was beyond the reach of mathematical analysis.

Koelreuter's concerns were limited, however: the boundaries imposed by nature on crosses between two plant species—"the most confused knot in the entire doctrine of reproduction." He did not

believe he could untie the knot, something for which human wisdom would perhaps prove too weak, but he did intend to set forth the obvious features that lay at its basis. Since in any one cross only two plant species were involved, Koelreuter conceived his problem as analogous to a combinatory. Specific differentia were susceptible of measure and enumeration; combinations would result in the addition of a sum or the subtraction of a difference. Koelreuter's tools, measurement and counting, were the same as those of contemporary systematists. The difference lay in the consistency with which Koelreuter applied these simple tools. His experiments with plant bastards were set up on such a scale and carried out in such a way that he could determine the different forms and sizes under which hybrid offspring appear, arrange these according to their different generations, and ascertain roughly some of their numerical relations. Koelreuter's results were products of the stubborn persistence with which he aimed at the strict numerical determination of carefully limited aspects of plant morphology.

I have mentioned results. Let me admit that I could no more have understood these results than did Koelreuter's contemporaries had it not been for the work of Robert Olby. In two short papers, Olby has made Koelreuter's work accessible as no previous historian has managed to do.^[2] However, Olby's insistence upon direct lines between Koelreuter and modern genetics is questionable and different from the approach taken here.

The Problem of Plant Sexuality

The competition on the question of plant sexuality proposed by the St. Petersburg Academy in 1759 provides an obvious point of entry into the study of plant hybrids and hybridization in the late 18th century. By the mid-18th century the question whether plants have sexual organs and reproduce sexually would seem to have been

decided by the many proofs offered by European naturalists. This was by no means the case, and many botanists continued to contest the idea.^[3] To end the conflict the Academy of Sciences in St. Petersburg proposed a prize question: "Either to confirm or to deny, by means of new proofs and experiments, as well as those already known, the doctrine of the sex of plants, preceded by a history and an account of all the parts of the plant that play some role in fructification and the formation of seed and fruit."^[4] The competition summed up received opinion on the subject of hybridization and initiated new research that marks the beginning of serious investigation of hereditary characters in plants. The Academy received three entries. Two were considered unsatisfactory; the third, however, was judged "praemio omnino dignus"; the prize was awarded to Carl von Linnæus, at the session of September 6, 1760.^[5]

Linnæus' paper offered nothing not already familiar to his readers; he simply restated positions on the subject of plant sexuality that he had defended throughout his career. In plants there is no fruit without flower; since the flower is a necessary antecedent to fruit, it follows that sexual organs must exist in the flower. In fact, Linnæus wrote, "the flower consists of nothing but sexual organs." Those parts with the rudiments of fruit, by analogy with the animal kingdom, must be female.^[6] Linnæus separated and combined male and female floral parts to show that seed developed only when pollen contacted the stigma, and argued that the possibility of fertilization of a female plant by a male of another species exhibited clearly the sexual duality of plants.^[7]

Linnæus considered the production of hybrids the decisive proof of the sexuality of plants. He cited four hybrids, but offered little by way of proof that they were the offspring of two separate species. One (Tragopogon hybridum ) had been hand-pollinated and marked

with a thread. (Koelreuter did not consider this cross above suspicion.) The other three plants (Veronica spuria, Delphinium hybridum, and Hieracium hybridum ), which had been found in the wild or in gardens, exhibited characters intermediate between two known species. Linnæus assumed without demonstration that these plants would reproduce through seed to form "constant varieties." This would constitute a new kind of metamorphosis in plants. "It cannot be doubted that we have here new species brought forth through hybridization."^[8] The bastard plant, although resembling the father outwardly, is the image of the mother with respect to the inner medullar substance and fructification. The numerous African Gerania , for example, led Linnæus to conclude that there are as many forms in one genus in the plant kingdom as have emerged from one species through the crossing of flowers. A genus is thus only the epitome of those plant forms that stem from a single mother and various fathers.^[9]

These speculations contradicted the few established facts concerning plant crosses. However plausible Linnæus' hypothesis, especially for genera in parts of the world where nature produces a copious variety of species, crosses between species are difficult to achieve. They are, moreover, impossible to perpetuate: because of their absolute or relative infertility, or because of the regressive degeneration to which their issue was subject, the aid of one of the parental stocks is required for fertilization. Insurmountable barriers of sterility, degeneration, and distribution limit the effects of crossing far more stringently than Linnæus imagined. Thus, for each species known to be distinct and constant, most naturalists assumed a common origin and epoch.

The key word is, of course, "assumed." The constancy of species characters and the absolute or relative sterility of crosses between species were subjects about which most naturalists of the 18th century found themselves obliged to make assumptions, just as they assumed that the development of a new individual was simply the

gradual distension of a preformed being. Trained in the identification and analysis of external conformation, they were at a loss to appreciate properties and novelties that breeding experiments alone might reveal. They classified so-called bastards as a special kind of anomaly, outside the rules ordinarily followed in reproduction.^[10] Each separate species is designed to function independently, yet to contribute "all those perfections towards the ends for which it had been determined."^[11] Intermediate forms lack this teleological justification, and naturalists reckoned among the wiser provisions of nature both the rarity of crosses between species in their free natural state and the infertility of bastards. Crosses between parents with very different organizations were seen as flatly impossible, and conjecture about crosses between genera and orders was idle speculation.^[12]

An adjunct of the St. Petersburg Academy, Joseph Gottlieb Koelreuter (1735–1806) had been a proponent of the sexual theory of plant reproduction since his early years as a student at Tübingen. One of his professors, J.G. Gmelin, had been among the first to recognize the significance of Linnæus' work on hybrids, and in 1749 had adopted the subject for his own inaugural lecture. In it, Gmelin called for experiments in hybridization.^[13] Koelreuter completed his

degree at Tübingen in 1755 and became an adjunct in natural history at St. Petersburg. When the Academy announced its competition on plant sexuality he recalled Gmelin's recommendation and set to work to produce plant hybrids.

Like Linnæus, Koelreuter considered the production of hybrids a decisive proof of the sexual duality of plants; unlike Linnæus, Koelreuter believed that nature limits this kind of anomaly, thereby preserving the order and harmony that had reigned in Eden. Two different species of animals living in a state of nature do not produce bastards; nature avoids disorder by means of natural instinct. She has equally certain methods for avoiding comparable disorder in plants.^[14]

Perhaps it has also been one of her intentions, in order to avoid just such a disturbing disorder, that she disposes one plant in Africa, and gives another its place in America. Perhaps it is partly on this account that she has confined within the limits of a certain region only those plants which in respect to structure have the least likeness with one another, and are consequently least likely to bring about disorder among themselves.

Bastards are products of the artifice of man, as exercised in botanical and zoological gardens.

Here at any rate man gives plants of a certain kind the opportunity that he gives his animals, often assembled from widely separated parts of the world, which he keeps penned in a zoological garden, or in an even narrower space.^[15]

Koelreuter was convinced that nature limits the potential for disorder, even under such unnatural conditions, and he set out to discover those limits.

Koelreuter produced his first plant bastards during the fall of 1760, after Linnæus won the Academy competition. The offspring of a cross between two tobacco species, Nicotiana paniculata and N. rustica , flowered the following March, and in the fall of 1761 Koelreuter published a brief account entitled Vorläufige Nachricht von einigen das Geschlecht der Pflanzen betreffenden Versuchen und Beobachtungen . He reported the results of continued experiments in

three Fortsetzungen (1763, 1764, and 1766). The Vorläufige Nachricht and the three Fortsetzungen offer a coherent account, not only of experiments in hybridization, but of the processes of pollination and fertilization. Koelreuter also published a number of individual papers on these subjects in the Commentarii of the St. Petersburg Academy.

The Vorläufige Nachricht continues the discussion occasioned by the prize competition. Indeed, the format of the piece follows closely the order prescribed in the prize question. For Koelreuter the production of plant bastards constitutes a decisive argument for sexual duality. Here, and only here, his analysis parallels that of Linnæus; elsewhere Koelreuter has taken pains to distance himself from Linnæus' wild claims and speculative flights.

The Combinatory of Parental Characters

Koelreuter's conception of hybrid production is inseparable from his account of normal generation. He rejected ovist and spermist theories of preformation; he also rejected Linnæus' theory of cortical and medullar layers—views he considered more clever than correct.

According to Koelreuter, two homogenous fluids of different kinds determined by the Creator for union with one another join to produce organization. The male agent in flowering plants is a product of the pollen grain; the female agent, Koelreuter believed initially, is a sticky secretion of the stigma.^[16] These fluids, male and female, differ essentially; that is, "the force of one must be different from the force of the other."^[17] From the union of these two fluids results another mean fluid, with a mean force compounded from the two simple forces. This purposive agent which emerges from the organization of the reproductive fluid as a whole is the source of organization for the future plant. For each class of organized beings, a specific compound force produces a determinate structure and specific nature.

Although Koelreuter made many observations and experiments concerning the process of fertilization, the mass of quantitative evidence he accumulated to support theories of epigenesis, equality of parental contributions, and sterility of interspecific offspring constitutes the novelty of his contribution to late 18th-century generational theory.^[18]

Koelreuter's work contained an important ambiguity concerning the source of organization in the living body. It is impossible to determine unequivocally whether the source of organization is material or nonmaterial.^[19] Koelreuter draws an analogy between the union of two seed materials and the production of salt crystals. When acid and alkaline substances unite, a third, intermediate salt results. In the same way, he argues, the intermediate fluid resulting from the union of male and female seed materials either constitutes the origin or the firm foundation of the vital machine, or produces this vital machine out of itself.^[20] Neither the male nor the female seed matter suffices to produce this result by itself, no more than an acid or alkaline substance can in and of itself produce the intermediate salt or form crystals. The formation of the plant requires both the compound of two specific seed materials and the composite active and purposive force resulting from that compound.^[21]

Nature works in the same way to produce a cross between two species. Once the male and female seed matters unite, formation proceeds rapidly from the nucleus to the flower, and in the process

the sharpest eye can find no more imperfection than in the natural plant. The resulting bastard, composed from the seed matter of two separate species, reflects twofold nature in both its intermediate form and its absolute or relative infertility.

When Koelreuter first worked to produce a cross between Nicotiana paniculata  and N. rustica , he experimented with many flowers. Each time his fertilization succeeded. The result was perfect, somewhat divergent seeds. From 110 seeds he produced 78 plants; of these, he kept 21 over the winter, and in March 1761 they flowered. In the spread of their branches, situation and color of the flower, and individual floral parts, each plant exhibited a mean between the two natural parental species.^[22] In repetitions of these experiments, the first-generation bastards consistently exhibited characteristics in "almost geometrical proportion" between differences in the parental species. Koelreuter tabulated his measurements:^[23]

Experiments with seven other genera supported the inference of intermediate size in first-generation hybrids. Koelreuter found that even the time of flowering and the odor of hybrid offspring were intermediate between the characteristics of the parents.

The intermediate character of first-generation hybrids, Kolereuter argued, supported the Aristotelian doctrine of reproduction by means of two seed matters. Koelreuter also used it to argue against the contemporary doctrine of generation, that is, "the doctrine of animalcula, or of original embryos and nuclei in the ovaries of animals and plants activated by male seed."^[24]

All offspring of the first cross, Nicotiana paniculata  × N. rustica , were identical. When Koelreuter reversed the direction of the cross, N. rustica  × N. paniculata , the offspring "agreed all together with the plants of the first experiment, and reacted in the same way to the experiments performed upon them."^[25] This phenomenon, now called the "identity of reciprocal crosses," also spoke to another contention about the contributions of parents to offspring. The hybrids with which naturalists were best acquainted, plant and animal, show a greater resemblance to the mother than to the father, leading naturalists to assume that this is a universal character of hybrids. Linnæus used the idea in his two-layer theory of generation: "a bastard offspring is with respect to its inner medullar [essential] substance the exact image of the mother, but in leaves and other outer [nonessential] parts [the image of] the father."^[26]

Koelreuter's reciprocal crosses and detailed measurements established, however, that parental contributions could not be distinguished so readily. Since Koelreuter did not come across any examples of sex-linked characters in his experimental plants, his inference that each parent contributed equally to an intermediate result was to an extent justifiable.^[27]

A second essential character of bastard offspring, fully as important as their intermediate form, is their absolute or relative infertility. Although the formation of the bastard tobacco parallels that of the natural species and its flowers are brilliant, the bastard plant is deficient in the most important character of all: fertility, the final cause of all formation. Koelreuter found the pollen containers of the Nicotiana bastard to be markedly smaller than in the natural species. What pollen they contained was white and dry; the grains did not cohere with one another as in the natural species. When examined under a microscope, the pollen grains proved to be irregular and shrunken. They contained scarcely any fluid, and most were empty husks.^[28] These observations led Koelreuter to doubt the fertility of the bastard plant; his experimental results reinforced this suspicion. Of the many flowers on the bastard plants, not one succeeded in bearing a single seed, even after dusting with a large quantity of the plant's own pollen: "instead of 50,000 [they] contained not a single one, and more than a thousand flowers, one after the other, fell without leaving a single capsule behind."^[29] In every sense of the word the tobacco proved to be "a true, and as far as I know, the first botanical mule produced by art."^[30]

Koelreuter thought infertility only a relative imperfection, which, from the point of view of ultimate consequences, proved to be a positive good. Nature wields sterility to preserve the order established at the Creation. "What an astonishing confusion would the peculiar and unchanged hybrid characters, and consistent fertility of such plants give rise to in nature? What a monstrous swarm of imperfections would they bear, and what evil and inevitable consequences would ensue?"^[31]

Koelreuter's pollination experiments, however, quickly convinced him that the infertility of the bastard was only relative. When pollinated with either of the parent species, the Nicotiana hybrid produced some ripe seed; much more was produced when plants of the

original natural species were self-pollinated.^[32] For each cross, Koelreuter raised ten plants from the bastard. The results were no longer intermediate between the two parent species, but measurably resembled the pollen parent. In Koelreuter's terminology, the plants were metamorphosed or transformed (his verb is verwandeln ): they again approached the fundamental nature (Grundwesen ) of which the initial crossing had deprived them.^[33] Again his measurements told the tale of transformation—in the size of plants; the spread of branches and flowers; and the form, size, and number of flowers.

Koelreuter's evidence for the relative infertility of hybrids and for the progressive reversion of hybrid offspring to the parental species parallels Georges Louis Leclerc Buffon's conjectures concerning animals published three years later. Under certain circumstances, Buffon maintained, a male mule can engender progeny and a female mule can conceive and give birth. Buffon also found a wide range in the productivity of different species of animals. This showed him that fertility is variable. Just as species vary in productivity, so, too, must hybrids—a conjecture that squared with Buffon's transformist views.^[34]

Koelreuter was less comfortable with the transformist position. Thus, when his experiments began to produce hybrids with varying degrees of fertility, he attempted to read the results as so many potential reversions to the parental species. He carefully constructed a table of his experimental results with respect to fertility. This was, he emphasized, not just another "useless, hasty, and absurd list of chimeric bastards,"^[35] but the first systematic catalogue according to a theory of generation certified by experiment.^[36] Koelreuter's class of "perfect" bastards includes offspring of two or three natural species of a single genus; normally they are infertile in the highest degree, although some products of only two species prove fertile on the female side when pollinated by either parent, or were fertile to a

diminished degree on both sides when crossed with parental species. "Imperfect" bastards are offspring of two natural species produced when a tincture of pollen from the female supplemented pollen from the male. These offspring are characterized by a diminished degree of fertility on both sides.^[37] Finally, what Koelreuter called "varietal" bastards are completely fertile. Koelreuter denied that the parents were different species. He drew a forceful conclusion:^[38]

Such a bastard in the true sense is either wholly infertile, or at most in a very limited and unequally diminished degree, by comparison with the true natural species from which it was produced is fertile. On the other hand, a mere bastard variety retains the degree of fertility of its parents, or at least loses nothing observable of this. Thus, I regard the experiment of crossing [species] in every respect as the only true, certain, and infallible touchstone of all separate species and varieties.

Koelreuter also distinguished between first- and second-generation hybrids. At an early stage in the experimental series, for example, he noticed a marked contrast in uniformity and stability. First-generation bastards for any single cross are all alike—intermediate in form between the parental species and either wholly or relatively sterile. Second-generation hybrids, even when produced from a single ovary, tend to be less like the parental bastard and more like the grandparents. After he fertilized perfect bastards of the first generation with pollen from Nicotiana paniculata or from N. rustica , he classified the results of such back crosses as descending (to the natural maternal species) or ascending (to the natural paternal species).^[39] Later he succeeded in producing second-generation hybrids from the self-pollination of the tobacco hybrid. Thus, contrary to expectation, pollen from the hybrid tobacco finally proved to have a slight degree of fertility. In Dianthus and Mirabilis bastards he found a greater degree of fertility on both sides.^[40]

Koelreuter used a quantitative argument to account for these results. Crosses in which the two seed matters are in equal proportion produce second-generation bastards resembling the first-generation parent; crosses of seed matters in unequal proportions produce second-generation bastards resembling more or less closely one or the other original natural species, depending on which seed matter is dominant and to what degree. These three main groups correspond to the three segregating classes of second-generation hybrids in Mendel's experiments with plants differing in one essential character.^[41] In other words, Koelreuter had found roughly the three segregating classes for second-generation hybrids. Koelreuter found puzzling, however, the exception to this neat scheme. Among the characters of second-generation Mirabilis bastards, for instance, he saw reversions, completely fresh characters, and colors too varied to be classified.

The contrasting results between first- and second-generation bastards required only a simple adaptation of the theory of normal reproduction. In any given plant the process of formation liberates the compound matter that gave rise to the form of the plant, and divides this compound once again into the two original matters, concentrated in the ovules and the pollen grains. In any natural fertilization the two seed matters and the simple forces inherent in them unite in equal proportions to form the intermediate product with its corresponding compound force, hypostatized during the process of generation in the specific characters of external form.^[42] In some original plant species the outer characters are very different, in others very similar; likewise, the seed matters and simple forces of natural species display differing degrees of affinity. Between species with little affinity—where characters and simple forces differ significantly—no crosses can take place. In cases of close affinity, where there is "no slight resemblance between its parents and a suitable agreement of their natures,"^[43] crosses can occur, although the Creator had not intended the two ground matters to unite.

Both vital functions in first-generation plant bastards differ from the norm: enhanced vegetative function and completely or partially curtailed reproductive function. Where infertility is not absolute, the next generation of offspring shows an unnaturally wide range of variability. The measurable intermediacy that characterizes all perfect bastards of the first generation characterizes only a small number of individuals in the next generation, the product of seed matters united in equal proportions. More often, however, mixture and union of seed matters do not proceed with the regularity typical of natural products and simple bastards. The principle of equality is broken; the seed matters combine in different proportions, and "All kinds of wrong paths result."

Two routes are open to nature according to Koelreuter. On one, "where she has the laws of close affinity as a guide, she again approaches the high road with something like a straight line; on the other path, where she lacks this guide, she strays . . . ever more from the high road."^[44] Here again Koelreuter's speculation parallels the best-informed contemporary opinion in zoology. In a discussion of sheep, goats, dogs, and domestic fowl the St. Petersburg academician Peter Simon Pallas noted that inconsistency of form, once introduced, increased from one generation to the next. Pallas supposed it to result from a vice introduced into the generative faculties of the original species by way of crosses. This vice, he said, parallels in effect the alteration of fluids and solids in a living body under the influence of a miasma.^[45]

The presence of essential characters resembling the original natural species was Koelreuter's key to the contrast between first- and second-generation hybrids. He concluded that although new varieties might arise through selfing, plant bastards do not establish the new and constant species Linnæus had predicted. Eventually hybrid races will revert to one or the other of the original natural species.

Offspring of Nicotiana rustica  × N. paniculata  are intermediate in form, produce sterile pollen, and are only somewhat fertile on

the female side. After pollination with N. paniculata , the plants grown from the seed resemble the pollen parent far more than the first generation. Koelreuter measured the essential likenesses with his usual care: situation, form, and substance of the leaves; number of leafless slender branches; shape and size of the corolla; form, color, and breadth of the flower; and form, size, and external perfection of the capsules.^[46]

The next summer (1762) Koelreuter pollinated this second generation with N. paniculata , and in 1763 he sowed 128 apparently fertile seeds, most of which germinated. Of this crop he retained 12 plants; when they bloomed they so resembled paniculata "that one could only have differentiated them with difficulty had they not been labeled with separate numbers."^[47] Moreover, the fertility of both pollen and seed had increased noticeably, although this was subject to fluctuations. Nicotiana paniculata was well on the way toward dominating rustica , and Koelreuter foresaw that this series of experiments would end by producing true N. paniculata . "In a word," he wrote, "I no longer have the least doubt of the possibility of transforming one natural species into another."^[48]

In the summer of 1765, four years after he had initiated this experimental series, Koelreuter achieved the first complete transformation of N. rustica into Nicotiana paniculata in the fifth generation, or, as he put it, in "the fourth ascending degree."^[49] It was as if "he had seen a cat emerge in the shape of a lion."^[50] A comparable reversion in the direction of the original mother plant, which he had followed in a parallel series of experiments, would, he predicted, require a different although proportional number of generations. He was also able to announce the successful transformation of Dianthus and Mirabilis . From a cross AB indistinguishable from the natural species B,

Koelreuter had thus produced offspring B. Since the maternal plant of AB was A, he reasoned that he had thereby transformed species A into species B. Moreover, he showed that the number of generations required for a successful transformation varied from species to species. He concluded that the transformation of "one plant into another [occurs according] to the greater or lesser degree of fertility [of] the bastards produced from their equality."^[51]

Transmutation of metals served as an analogy for this experimental series. In theory one metal could be transformed into another by taking from it particular properties, conceived as so many independent substances, and substituting other properties for them. Mercury, for example, could be ennobled by removing the two characters upon which its fluidity and volatility rested and substituting other characters for them. In the process two seed matters were needed: the male agent, of a sulfuric nature, possessed the force to make the fluid mercurial female seed matter capable of resisting fire and forming a stable body. Sulfur, in other words, transformed the nature of the mercurial body. The result was a true metamorphosis in which the male, sulfuric agent asserted its superiority (Uebergewicht ) over the mercurial agent.

Koelreuter believed he had achieved the alchemist's dream in botany in the course of a few years. In plants the male seed matter is oily and sulfuric; union with the female seed matter produces a stable organic body, "the initial basis of the future plant." In succeeding generations the male agent gradually takes the upper hand, and in the end the nature of the female has been wholly transformed. The transformation depends on "close affinity, fertility not wholly suppressed in the production of bastards, and dominance in a certain degree."^[52]

Koelreuter used his success to argue for the possibility of the transformation of metals. Likewise, one species of animal can probably be transformed into another. From a canary might come a linnet. Experiment had already established that the female retains

fertility in "the second descending degree," and it was therefore probable that she would do so in the "second ascending degree."^[53]

Central in Koelreuter's reasoning was the contention that a measurable change of characters in the product of any unnatural union reflects the proportions of the two seed matters. Variability increases in the offspring of bastards since the number of possible combinations of the two seed matters is infinite: that thread of Ariadne, natural affinity, is lacking. Koelreuter did see limits to variation, however. By means of crosses a naturalist can transform one affine species into another, or produce one of the infinite number of possible variations between the forms of the two natural species, but cannot produce wholly new species with entirely new characters. The naturalist thus transforms natural bodies by altering or removing particular characters and substituting others.

Transformations of this kind refuted monoparental heredity, and they showed that crossing was a powerful instrument for change in the world of living forms. But how was change itself interpreted? Koelreuter read his experiments against a conceptual tradition that did not permit development away from the original natural forms. He considered bastard plants artificial products, able to survive only in artificial conditions. "In the orderly arrangement and ordinary situation," he wrote, "established by nature in the plant kingdom, bastard plants would be difficult to produce or even to initiate." Even if we suppose the possibility of a true bastard plant in an open field, "the question would remain whether this chance had not taken place in a region where the natural situation as a whole, either mediately or immediately, had been destroyed or changed: true wilderness as it comes from the hand of nature is one thing; a field, free but in respect to a hundred things often very much altered by the hand of man, is another."^[54]

Consider Verbascum , for which the probability of natural crosses is great. Koelreuter did not believe that unnatural products could transform the natural species, even when, as in Verbascum , the

natural species cross regularly. In an open field, where its own and alien pollen reach the stigmata at approximately the same time, a plant will accept the male agent intended for it by nature, excluding the alien matter from fertilization. Neither ancients nor moderns spoke of bastard Verbascum in the field. Linnæus, it is true, mentioned a Verbascum bastard, but this was likely a product of the unnatural conditions in the Uppsala botanical garden. Koelreuter continued, "It is to be wished, nonetheless, that Herr von Linné had given us a more careful description, and more according to nature than according to his fantastic theory of generation, which contradicts nature."^[55]

As a naturalist, Koelreuter may have sensed the limits and assumptions of his conceptual tradition, but his was a stubborn defense of the traditional essentialist concept of a species. Koelreuter used his experiments to throw light on the nature of the species as a whole, not the inheritance of individual characters. He set out to prove that a cross between two species does not produce new and constant species. We may imagine that his findings were in this respect a relief. He regarded the essence of species as monolithic; the intermediacy that characterized all first generation bastards confirmed this interpretation.^[56] However, second-generation and back-crossed hybrids did not lend themselves to the same neat explanations. Here Koelreuter took comfort from the predominance of forms resembling the original natural species; from this he concluded that hybrid offspring will revert sooner or later to one or the other of the original species. They cannot, he thought, from new and constant species.

Reception

Contemporaries were only too happy to accept Koelreuter's conclusions. The rigor and complexity of his experiments were alien to

the mentality of many naturalists of the late 18th century, but his results supported very powerfully the theory of limited variability then being elaborated. Nor was Koelreuter's work so obscure or inaccessible that no one read it. Christian Sprengel took up Koelreuter's work on pollination, arguing that floral structure was an adaptation to secure pollination and showing that cross-fertilization was the rule rather than the exception. C.F. von Gärtner, an erudite and industrious plant hybridizer reaffirmed Koelreuter's entire theory of generation. His evidence consisted of the results of nearly 10,000 separate crossing experiments among 700 species. As Koelreuter's Vorläufige Nachricht and the three Fortsetzungen were published, a reviewer in the Göttingsche Anzeigen astutely picked out the tendency of the argument. Not only had Koelreuter limited severely the possibility of novelty by means of species crosses; he had shown that change was a narrowly defined combinatory of already existing characters.^[57]

These summaries, and mémoires published by Koelreuter between 1770 and 1796 in the Commentarii and Acta of St. Petersburg, were read, assimilated, and commented upon by two of the most influential and widely read naturalists of the period, Johann Friedrich Blumenbach and Peter Simon Pallas. In his first important work on the theory of generation, Blumenbach presented Koelreuter's experiments as a refutation of preformation. The very possibility of specific bastards contradicted any conception of preformed germs, while Koelreuter's transformation of one species into another "must undeceive even the most partisan defender of the theory of evolution [i.e., preformationist] from his prejudice."^[58]

Peter Simon Pallas claimed that Koelreuter had instituted "experiments on plant crosses the results of which absolutely contradict the opinion of the Chevalier Linné." He continued:^[59]

The great difficulty we have in producing crosses, with all our human industry, between two different species, the impossibility, confirmed by fact, of perpetuating these as crosses or as distinct races, either because of their

absolute or relative infertility, or because of the regressive degeneration to which their issue is subject, since these crosses require for fertilization the aid of one of their primitive branches, all this opposes facts to a simple probability and forces us to see that all those species that nature affects to render alike as primitive, projected in the first plan of creation, and destined to form that chain of beings that we admire without being able to account for it, any more than [we can account for] the choice, harmony, and combination of colors and ornaments that the same creative force has used to embellish its works.

The experiments Koelreuter undertook to discover the reproductive limits of plant species may seem commonplace. Older natural history had a modest experimental tradition and breeding experiments utilizing plants and animals were common after 1750. Koelreuter was unusual, however, in his diligence and energy: his experiments involved 138 species and over 500 hybridizations, and required him to examine pollen grains from more than 1,000 different species of plants.^[60] No other experimental sequence in natural history equals Koelreuter's in extent.

Koelreuter's use of counting and measurement in the analysis of his data was also unusual in 18th-century natural history. Although his contemporaries were masters of impersonal, objective, precise observation, naturalists throughout the 18th century ignored or resisted the use of mathematical methods and symbols in their studies of morphobiology. Natural history, as they understood it, was observational, qualitative, and finalist. The central conceptual tradition was Aristotelian, with its depreciation of mathematics as irreconciable with the study of final causes.^[61] Thus a resistance to measuring and counting came naturally to 18th-century natural history.

Not all naturalists were Aristotelians, however, and even the diehards of the Aristotelian tradition found it impossible to resist the forces toward greater precision that originated in more prestigious spheres of the scientific constellation. Anatomy contributed to the establishment of species characters the procedure of measuring size. Naturalists also counted seeds and pollen grains, observed numerical

ratios, noted numbers of progeny, and proposed mathematical nomenclatures. These efforts toward quantification, however, remained sporadic and uncorrelated; and most naturalists considered the complexity of living forms beyond the power of mathematical analysis.

Koelreuter limited his concern to the boundaries imposed by nature upon crosses between two plant species—"the most confused knot in the entire doctrine of reproduction." He did not intend to untie the knot, but rather to identify its major features. Since only two plant species were involved in any one cross, Koelreuter conceived his problem as analogous to a combinatory. Because specific differentia were susceptible to measure and enumeration, combinations could thus be expressed as a simple sum or difference. Koelreuter applied the instruments of measurement and counting with unusual consistency. The design and scale of his experiments permitted him to determine different forms and sizes of hybrid offspring, arrange them by generation, and estimate their ratios. That his quantitative approach yielded important results was due in large measure to his stubborn concentration on carefully limited problems of plant morphology.