Chapter 10
Analogical Reasoning:
The Theory

The hypothesis that sap circulates in plants just as blood circulates in animals was not the first attempt to liken plants to animals. Animals had served as models for explaining plants since ancient times, and seventeenth-century savants equated seeds to eggs, named the parts of plants after the parts of animals, and compared the structures of plants and animals.^[1] Such ideas became productive in the late seventeenth century, however, because botanists relinquished the Aristotelian distinction between animal and vegetative souls, which separated flora and fauna into separate causal categories, and because they sought causal mechanisms instead of causative "faculties."^[2] What had formerly been lazy analogical thought that merely affirmed the unity of living things became instead a guide to the empirical investigation of causal mechanisms. Thus, Harvey's theory reanimated botany at the end of the seventeenth century because it offered a heuristic model: it suggested experiments and practical applications that might ensue from a transplanted theory, and it stimulated the search for a causal mechanism.^[3]

The Circulation of the Sap

In the summer of 1668 Claude Perrault and Edme Mariotte defended the hypothesis that sap circulates in plants as blood does in animals. They identified five ways in which plants resemble animals: plants have two sorts of vessels, corresponding to veins and arteries; there are two sorts of sap,

and these are the equivalents of venous and arterial blood; sap is nutritious for the plant, just as blood nourishes the animal; the root manufactures sap just as the liver produces blood; and sap circulates frequently and quickly through the plant, replenishing itself with water from the leaves and being recooked in the root, just as blood circulates and is refreshed during its circuit through the body. Like Harvey, they had a hierarchical concept of the organism and emphasized "the control and stewardship" of one part of the body over the others, and like Riolan they incorporated traditional physiology — for example, the idea that the blood was itself a nutriment — into their theory.^[4]

The debate that the Academy sponsored marks the first systematic effort to apply circulatory theory to plants. Similar ideas were current outside the Academy in the 166Os: Johann Daniel Major had suggested the analogy, Timothy Clark had written about a circulation of the liquid in sensitive plants and had searched with a microscope for structural equivalents of valves, and Nicaise Le Febvre had compared the functions of sap and blood. In the 1670s and 1680s Nehemiah Grew and Marcello Malpighi impressed the botanical world with their systematic studies of plant anatomy and physiology.^[5] But it was Mariotte and Perrault who first pushed the analogy between blood and sap to its limits.

The debate of 1668 represents one of the Academy's most productive efforts at refereeing research. It began with the conflicting claims of Mariotte and Perrault for priority, and ended amicably by recognizing that their independent judgments had coincided. The exchange of evidence and opinions in 1668 influenced not only the two protagonists but also Duclos. Originally drawn into the debate to review the evidence, Duclos opposed the theory during the summer of 1668 but supported it in 1680. Finally, Perrault, Mariotte, and Duclos published their views about the theory.

The essential scientific traits of Perrault and Mariotte are exemplified by their work on the circulation of sap. Perrault was theoretical. He conjectured, offered plausible arguments, and identified the need for experimental support. In citing experiments, however, he rarely used the first person, and all the experiments cited in his 1680 book were actually performed by Mariotte, Huygens, Duclos, and Bourdelin at the Academy, not by Perrault himself.^[6] In contrast, Mariotte was emphatically experimental, and even his initial inspiration that sap circulates was prompted by an experiment.^[7]

Proving the circulatory hypothesis required academicians to address the three issues that Hesse has identified as crucial. Academicians had to establish the pretheoretic similarities between plants and animals that would make the analogy materially plausible; here they relied on lazy

analogies and on a functional resemblance. Next they had to push the analogy to its limits, testing for traits in plants that would correspond to those in animals; here academicians identified crucial dissimilarities that ruled out the relations of causality they originally anticipated. Finally, they were faced with the problem of crucial dissimilarities: plants were not comparable to animals in several significant respects, and in particular they lacked any internal motive force that could pump the sap as the heart pumped the blood. Because this latter dissimilarity could not be resolved, the analogy as a whole failed, and Mariotte and Perrault were left with a most difficult and important botanical question: how does sap rise in the first place? For an answer they turned ultimately to disciplines other than botany or zoology.

Pretheoretic Plausibility

Botanists in the Academy first had to show that the circulation of sap was plausible. Like Harvey, who reminded his readers of other circular motions in nature and claimed "as much right to call this movement of the blood circular as Aristotle had to say that the air and water emulate the circular motion of the heavenly bodies," Perrault compared the circulation of sap to the cycle of condensation and evaporation. Harvey and Perrault both dwelt on the differences between living creatures and other things, Harvey stressing the connection between heat and life, Perrault the "natural connections" that unite the various parts of the body. Mariotte emphasized another of Harvey's themes, the connection between movement and life. Harvey affirmed that motion is necessary to generate the heat that is associated with life and pointed out that blood clots when it does not move, and Mariotte observed that still liquids stagnate and become corrupt or death-like.^[8]

Pretheoretic analogy or plausibility rested here on lazy analogy. Plants and animals were living creatures that depended for life on the motion of liquids that transported vital heat to all their parts. These were seventeenth-century truisms about the nature of life. Better pretheoretic support came from the specific context in which circulation of the sap was proposed. Perrault and Mariotte believed that this was above all a question of nutrition.

In 1667 Perrault had introduced the theory of circulation as a way of explaining how plants are nourished. In the following year he cast his entire analysis in the context of whether plants and animals are nourished similarly. When Mariotte explained what aspects of blood he meant to compare with sap, he started with the reception of chyle by the lacteal veins in the

mesentery and their transmission of this food to the venous blood. Following this non-Harveian interpretation of the lacteal veins,^[9] Mariotte described the full circuit of the blood and then a hypothetical circuit of the sap:

Probably the ends of the roots imbibe liquid from the earth and carry it into the body of the root. From there it passes into small vessels in the stem; and then it is distributed to the branches and ends of the leaves. The remainder is carried along different small channels to the root to be perfected by a type of cohobation and in order to become a well-digested sap, appropriate for the nourishment of flowers and fruits.^[10]

When Mariotte published the theory, he called his treatise "On the Vegetation of Plants," and he embedded the circulatory theory in a broader discussion of the chemical composition of plants, their germination and growth, the origins of vegetable nutrients, and the effects of plants on other living creatures.

These were the general grounds on which circular motion of sap was plausible. In addition to the vague principle that motion is necessary to generate life-sustaining functions, academicians cited the more specific need for nourishment characteristic of all plants and animals. A functional rather than a formal resemblance between plants and animals stimulated the analogy. But to confirm it Mariotte and Perrault still had to show that sap actually circulated and to find structures in plants that resembled the circulatory organs of animals.

Pushing the Analogy to Its Limits

Academicians sought to prove that sap circulated, first, by showing that their general theory of growth necessitated circulation and, second, by developing experimental evidence that demonstrated the descent of sap. Scientists at the end of the seventeenth century explained growth mechanically as resulting from the pressure of blood or sap against vessels in the extremities of animals and plants. That is why, explained Perrault, the French use the word "pousser" to speak of growth. Since all parts of plants and animals grow larger, all must be subject to this pressure. From a mechanistic theory of growth, it followed that sap must push downward as well as upward, because roots and the tops of plants grew, downwards and upwards respectively, in proportion to one another.^[11]

Experimental evidence for circulation was sought in several ways. Everyone assumed that sap rose from the root to the top of the plant. The

novelty was in showing that it descended again to the root. In 1668 Mariotte had already performed most of the experiments that proved descent of sap from the tops of plants toward the roots. He cut stems and observed that sap flowed in both directions. When he planted seeds upside-down, or placed them with the leaf-end in water and the root-end exposed above the water, the seeds grew. When he cut the filaments on roots, they bled. Uprooted chives, placed in water with only the shoots immersed, survived and grew for a fortnight. Most of the early experiments immersed seeds, plants, or parts of plants upside-down in water. Usually the plant survived and grew for several days. From these experiments, Mariotte and Perrault concluded that sap did move toward the root, that they had found two kinds of sap (yellowish and whitish in color, or thick or thin in consistency) in most plants, and that leaves could absorb water.^[12] Proving experimentally that sap descends to the root hence led to a promising observation, that there were two kinds of sap, and to an unforeseen consequence, that leaves had an important role in the nutrition of plants.

If leaves could absorb food, they endangered the analogy with animals, because an animal ingested food through a single mouth. Mariotte tested leaves in water and observed the sap in their branches. He concluded that leaves not only absorbed water but also carried more water than did the root to the vessels containing yellow sap. He also noticed that drops of water formed on the leaves of plants under a glass bell. Assuming that these were dewdrops, seeing that the plants remained healthy, and believing that no other water was available to the plants, Mariotte concluded that the leaves absorbed enough water to sustain the plant.^[13] Mariotte, the experimentalist, showed that plants, unlike animals, could take their food through two orifices. Perrault, the theoretician, tried to reconcile the new evidence to the analogy with animals. Here Harvey's own theory offered an explanatory model, for he had noted that medicaments applied externally to one part of the body entered the bloodstream and traveled to the entire body. Perrault, therefore, suggested that leaves absorbed liquids the way the skin of a dog absorbs the heat of a fire, or the skin of a butcher the fat of the meats he is handling. Mariotte affirmed the consequences of the experiments, while Perrault drew on folk wisdom and lazy analogy to reconcile the behavior of plants and animals.^[14]

If the descent of sap proved the circulation of sap, the existence of two kinds of sap in a plant provided another likeness between blood and sap. Harvey had shown that there were two kinds of blood: one, going out to the body from the heart, that was "hotter, perfect, vaporous, spirituous and, so to speak, nutritious," and one whose nourishment and heat were exhausted

and which returned to the heart "cooled, coagulated, and … figuratively worn out."^[15] For sap to be comparable to blood, therefore, rising sap should nourish and falling sap be weak and useless. Perrault asked whether there was a thick sap equivalent to arterial blood and a watery sap equivalent to venous blood, and Mariotte found plants that contained two different saps, yellow or white, thick or thin.^[16] The difficulty was to show that one was more nutritious than the other. Citing trees tapped for their sap in the spring, Perrault argued that falling sap cannot be nutritious, or trees would die from the loss of so much of it. Academicians analyzed the saps chemically but could not agree on the results.^[17] Although academicians compared the two sorts of sap with venous and arterial blood, they asserted resemblances more effectively than they proved them.

The analogy also constrained academicians to find separate vessels for carrying sap up and down the plant. In 1668 Mariotte could not determine whether there were two kinds of vessels. Perhaps mindful of Harvey's insistence to Riolan that one vessel could not accommodate simultaneously the flow of two liquids in opposite directions, Mariotte inferred from the two kinds of sap that there must be two sorts of vessels. Perrault, on the contrary, defended Riolan's stance, and tried to show by analogy that a vessel might permit both upward and downward flow, if the two liquids were sufficiently different.^[18] By 1679, Mariotte had studied the anatomy of plants more systematically and with a magnifying glass. He described the appearance of stems and the arrangement of fibers, channels, and spongy matter in them.^[19] But he still could not positively identify different vessels for different saps. In 1680, Perrault argued from Mariotte's 1668 experiments and from general information about trees that some plants had separate vessels and were therefore like "perfect" or higher animals. But he cautioned that the absence of separate vessels could not disprove circulation. Some plants were like insects, which do not have separate vessels. Perrault argued that the parts of plants must be able to differentiate cooked from raw sap, perhaps because of the disposition of their pores; in other plants the double bark or the bark and the pith might serve as separate conduits for different saps.^[20]

Equally serious was the lack of a mechanism for controlling the direction of flow. Mariotte and Perrault were uncertain whether they could identify valves or equivalents that prevented the rising sap from falling prematurely.^[21] Among academicians, La Hire agreed with Robert Hooke's sentiment that valves seemed "very necessary for conveying the juice of trees up to the height of sometimes 200, 300, and more feet; which he saw not how it was possible to be performed without valves as well as motion." La

Hire claimed to have found valves in canes and reeds.^[22] Allowing expectation to prejudice observation, he argued from similarity of function to resemblance of structure and insisted that he had found valves where none existed.

Skepticism about the existence of valves or separate conduits required academicians to consider whether sap flowed in different directions in the same vessel. Grew believed this was so.^[23] Perrault tried to explain that it was possible by using an analogy with water vapor, which can rise through oiled paper but cannot penetrate it again once the vapor has condensed. He offered a second analogy, with two sponges, one soaked in water and the other in oil, each of which absorbed liquid of its own type when placed in a mixture of oil and water. Both arguments presumed that the parts of plants were designed to accept one kind of sap and reject the other.

Each structural comparison endangered the analogy. The closest resemblance was between blood and sap, liquids that were of two kinds and that made a complete circuit of the body in question. Organs and blood vessels, however, challenged the ingenuity of savants. Academicians tried to compare the root to the heart, the soil in which a plant stood to the intestines, but even with a microscope they could not identify two distinct sets of vessels. The most serious difficulty, however, was that since plants lacked a heart, they had no pump to drive the sap.

Solving the Problem of Crucial Dissimilarities

Failure to find equivalent organs weakened the analogy seriously. The absence of what Hesse calls horizontal analogues meant that the causal mechanism was missing in plants. Academicians were reluctant to discard the circulatory hypothesis altogether, once they had found evidence that sap descends. But purported equivalents, putative valves, and two-way vessels were inadequate to explain the principal problem of the circulatory analogy, the very issue Perrault had identified with his opening words in 1668: how does sap rise in the first place?

One escape from the failed analogy was to compare plants with "lower" animals, whose anatomical organization was less differentiated. Such a move was consistent with the Harveian model, for Harvey had noted that valves are "not present in all animals" and are not "made with equal skill in all the animals in which they are present." Harvey had also allowed for circulation in animals that lacked hearts, and plants seemed similar to what Harvey had called "plant-animals." These were creatures such as oysters

and earthworms, which had only rudimentary hearts or no hearts at all, because they were too small, too cold, too soft, and "too little differentiated in their structure." Harvey admitted that they did not need "a propulsive organ to transmit food to their extremities." Their bodies were limbless and homogeneous. In them, ingestion "and expulsion of food is an in and out movement produced by contraction and relaxation of the body as a whole." They need no heart because "they use the whole of the body as such and an animal of this sort is in effect nothing but a heart."^[24]

Valveless veins and heartless creatures, therefore, offered two escape routes to scientists who used Harvey's theory as a model for plants. A heart, valved veins, and pulsating arteries were not necessary for circulation in all animals, since the entire organism might serve as a propulsive mechanism, driving nutrients and excrement through itself. But plants differed from the plant-animals: they tended to be larger, had a more differentiated structure, and were stationary. Botanists who took the view that the entire plant could propel the sap upward would have to look for an external motive force. To show how sap rose at all, therefore, academicians had to look beyond anatomical or structural identities. They had to seek nonbiological forces operating outside or within the plant. Botanists still sought a pump, not organic but figurative, that could impel sap in a direction contrary to its natural downward flow.

In the absence of a biological causal mechanism, academicians turned to two explanatory modes: chemical and physical. The former operated inside the plant and was part of the normal physiology of a living creature. The latter could be either external or internal, depending on what kind of phenomenon was cited. Chemical explanation was consistent with the Harveian model, for the digestion of nutrients was thought to be a chemical process that resulted in effervescence and rarefaction of the digested substances. The physical explanation, on the contrary, relied on concepts of air pressure and capillary action unknown to Harvey. Both chemical and physical mechanisms were invoked by Perrault and Mariotte, who thereby diluted the biological model with nonbiological explanatory mechanisms. But the physical model itself was uncertain, since scientists in the late seventeenth century were unclear about the causes of capillary action. Thus the analogy with the circulation of blood, which had a visible and organic causal mechanism, led botanists to adopt as a causal mechanism a mysterious phenomenon that had only recently been investigated. When the analogy between the motions of blood and sap could not be sustained, academicians related the rise of sap to capillary action and air pressure.

When those explanations seemed inadequate they cited the chemical phenomena of effervescence and rarefaction.

Explaining the Rise of Sap

The concept of capillary action was regarded as novel during the 1660s. Robert Hooke believed the phenomenon had first been observed by French scientists:

An Eminent mathematician told me one day, that some inquisitive Frenchmen (whose Names I know not) had observed, that in case one end of a slender and perforated Pipe of Glass, … be dipt in water, … the liquor will ascend to some height in the Pipe … tho held perpendicular to the plain of water. And to satisfie me, that he mis-related not the Experiment he soon after brought two or three small Pipes of Glass, which gave me the opportunity of trying it.^[25]

Robert Boyle experimented with capillary tubes in 1660, discussed capillary action in 1671 and 1676, and tried to measure the force of capillary imbibition in a seed. Huygens knew of his work, saw demonstrations of capillary action at Rohault's house in December 1660, and attended a meeting at Gresham College in April 1661 where the phenomenon was discussed; he also owned a copy of "Boyle's article on the rise of water in small tubes and on other phenomena which we call capillary."^[26]

Capillary action seemed to seventeenth-century scientists to explain several natural phenomena. Hooke listed these effects of capillary action:

… the Rising of Liquors in a Filtre, the rising of Spirit of Wine, Oyl, Melted Tallow, & c. in the Weake of a Lamp (tho made of small wire, threeds of Asbestus, Strings of Glass, or the like) the Rising of Liquors in a Spunge, piece of Bread, perhaps also the ascending of Sap in Trees and Plants, through their small, and some of them imperceptible Pores, (of which perhaps I may say more on another occasion) at least the passing of it out of the earth into their roots.^[27]

He believed that capillary action was the result of unequal air pressure, with the air pressing heavier on the reservoir of water surrounding the thin glass tube than on that within the tube itself. Hooke reiterated the view in 1665 and stated that air pressure caused sap to rise in plants, basing his explanation on the analogy with capillary action.^[28] This confusion between capillary action and the effects of air pressure persisted throughout the century, despite G. A. Borelli's argument in 1670 that capillary action was not due to air pressure.^[29]

Because capillary tubes resembled the stems and vessels of plants, they were an obvious model. In the late 1670s Mariotte and La Hire used the analogy between glass tubes and the vessels of plants to explain how sap rose. Mariotte limited the effect of capillary action, however, to "the first entry of the water into the roots." This, he said "occurs by a law of nature similar to the movement of union of which I have already spoken; since whenever very narrow tubes touch water, it enters, and it even rises despite its natural tendency to descend." For capillary action to work in plants, three conditions had to exist: the water in the soil had to touch the plant, it had to have access to that part of the plant in which it could rise, and something had to cause the liquid to enter and rise in the plant. Mere contiguity seemed to be insufficient because if a glass tube were dirty or rubbed with tallow, water would not rise in it. Furthermore, the pores of plants had to be properly "disposed to allow the subtle parts of other bodies to enter." Finally, these subtle parts had to "be pushed by some principle of motion." This principle Mariotte referred to as "a movement of union" and as an effect "that is popularly called attraction"; he did not cite the pressure of air as Hooke had done.^[30]

But there were problems inherent in comparing the rise of sap in plants to the rise of liquids in capillary tubes: the cause of capillary action was disputed and liquids did not rise so high in capillary tubes as they did in plants. Thus, whether or not capillary action was cited, savants turned to air pressure in order to explain the rise of sap in taller plants.

Huygens initially favored the view that sap rose because of air pressure, and in 1668 Perrault propounded an explanation that depended on multiple causes, including air pressure.^[31] The theory that air pressure causes sap to rise was known to be defective, however, by 1679. Pierre Perrault and Huygens debated whether sap rose because of air pressure, as Huygens maintained, or because of attraction and nature's abhorrence of a vacuum, Perrault's view. Perrault cited the following as a decisive argument against Huygens's theory:

How can we understand the sap that rises in trees? Can one say that air pressure causes it to rise between the bark and the wood, as in a pump? For that it would be necessary for the foot of the tree to rest in a reservoir of sap. Even if that were the case, this sap could rise only thirty-two feet, but there are trees that are more than one hundred and twenty feet tall.^[32]

Pierre Perrault concluded that sap rose as a result of "attraction due to abhorrence of the void," a phrase that he and Mariotte used apologetically

and as a manner of speaking not intended to impute emotions to inanimate or mechanical objects.

Since neither capillarity nor air pressure seemed sufficient to raise sap higher than thirty feet, some savants adduced chemical reactions inside the plant. Claude Perrault had argued all along that sap rose for many reasons, including the existence of appropriate passages in the plant, air pressure, external propulsion from the wind, and coction of sap. Influenced by his brother's objections, however, he modified this view when he published his Circulation in 1680. There he cited both air pressure and attraction due to fear of the void, but he also elaborated in Cartesian fashion his earlier chemical explanation: when sap was prepared, fermentation and effervescence reduced its concentration so that more sap rushed in to fill the potential void. To show that sap would flow into an area of diminished pressure within a plant, Perrault cited the effects of an air pump: "plants that are full of sap let it run when the air is being evacuated, and when the pressure of the air is diminished, the sap dilates and becomes less condensed than it was."^[33] Mariotte developed a slightly different theory in 1681: sunlight evaporated sap in the upper regions of the plant, creating an area of lower pressure into which sap rose.^[34] Such formulations anticipated Hales's conjecture that the loss of liquid due to transpiration pulled sap upward, continuing a process begun by capillary action.^[35]

Tournefort presented a different eclectic theory in 1691. He believed he had found two different systems in plants, one in which sap rose by absorption and another in which it rose by capillary action. The vessels in most plants, he said, were soft, spongy, and composed of many small, empty bladders or pouches that were connected so that sap passed through them. He compared them to felt strips or cotton that filtered and conducted liquids. Not all vessels were spongy, however; the stems of water-plants were like cylinders pierced longitudinally with holes. These tubes carried sap, and Tournefort thought they resembled capillary tubes: "this structure seems to favor the sentiment of some natural philosophers who believe that the sap ascends in plants for the same reason that water rises in very narrow glass tubes."^[36] Tournefort followed G. A. Borelli, who claimed that the dilation and condensation of air enclosed in plants caused sap to rise and that the spongy matter in plants facilitated that rise.^[37] La Hire, however, challenged the Borelli-Tournefort hypothesis after observing that water did not rise significantly in absorbent materials such as sponges inserted in glass tubes or paper strips. The best result was a height of 225 lignes over a period of more than eighty-four days.^[38] La Hire concluded that neither absorbent matter nor capillary action could account for the rise of sap.

Instead, he maintained that only hollow tubes in plants could transport sap, and he claimed to have found hinged, woody valves in them that enabled sap to rise.^[39]

The problem of how sap rose elicited varied responses from academicians. Huygens proposed air pressure, while Perrault listed several interrelated causes, such as air pressure, wind, fermentation, and the different weights of raw and cooked sap. Once air pressure was ruled out as a sufficient cause, Mariotte and La Hire focused on capillary action, while Tournefort combined capillary action with a Borellian theory about spongy matter in plants. Later, La Hire adopted the view that the vessels of plants were valved. In each case, academicians used analogies. Since the biological model of valves was inapplicable, they drew mostly on chemical or physical explanations. But the limitations of all models forced most academicians to develop theories based on multiple causation.

Conclusion

Academicians used the hypothesis of a circulation of sap to search for causal mechanisms. By choosing the Harveian model botanists tried to replace the two principal modes of explaining living things — chemical and mechanical — with a biological one. When the analogy failed, academicians had three choices. They could force a biological explanation by insisting on false structural resemblances. They could fall back on either or both of the traditional modes of explanation. Or they could draw on new physical models that were only half understood. Since the Harveian model itself assumed chemical processes within the physiological and retained a technological model for the heart, recourse to nonbiological explanation was broadly consistent with the model.

Circulatory theory had both substantive and methodological shortcomings. Although a circuit of sap was established, structural resemblances alone between plants and animals could not justify a causal analogue. The effort at methodological equivalence fared no better, despite the experimental ingenuity displayed by Mariotte and La Hire, whose demonstrations of the direction of flow sometimes resemble Harvey's tests. Harvey's crucial experiments were done on living creatures, however, and his theory owed a great debt to his skill at vivisection, but all of Mariotte's and La Hire's dissections were of dead plants. Experiment is necessarily an artificial procedure that may distort its object, and it is least informative when it examines defunct organisms in order to understand physiological processes.^[40]

Given these failures of the analogy, did it have any value at all for seventeenth-century botanists? First, the circulatory analogy had useful consequences despite being a weak form of a partially failed induction. In the absence of a compelling alternative, botanists found a partial analogy better than none at all. Although the causal resemblance failed, a circuit of sap was established. By calling that a circulation, botanists implied that plants enjoyed the digestive and perfecting processes characteristic of animal nutrition. While the circulation of sap could not be subsumed under the laws governing the circulation of blood, the term "circulation" reminded botanists of both the circuit and its function, if not its causal mechanism. Hence, the analogy with the movement of blood supplied a suggestive language to botanists, who retained some of the connotations that the word "circulation" had grown to have for natural philosophers.

Second, because academicians used the experimental and observational form of analogy, their analogy was both positively and negatively useful. It aided "the investigation of structure and the relation of structure and function" and helped reveal some "properties hitherto unnoticed."^[41] It also led academicians to ask new questions and to propose different causes, causes that were testable. Because the extension of the metaphor was more limited structurally than had been foreseen, however, botanical investigation did not affect notions about the physiologies of animals or "plant-animals." Finally, those who did not insist on putative valves in plant vessels learned from the circulatory analogy that there was no organ in plants to make sap rise.

Third, when a model cannot lend its mechanism because of structural disanalogies, the principal value of analogical reasoning must be as an inducement to comparative method. In the case of plants and animals, this may be inevitable, for the two types of organisms enjoy similar functions but dissimilar structures. Analogies used experimentally can draw attention to these problems. But for analogy to work as comparative method, the researcher must not assume the identity of the two things being compared. Therefore, an analogical argument must start by elucidating similarities and differences, as Hesse has pointed out. In this case, the analogy with animals forced botanists to ask how plants accomplish certain functions without having the appropriate organs. By retaining the analogy while admitting structural dissimilarity, academicians moved from analogical to comparative method. That is, they used the analogy to locate specific resemblances and differences; they then tried to explain the differences by comparing the causal mechanisms of the two. In the case of the Academy's study of sap, this transition was incomplete. Some academicians

persisted in pressing the structural comparison by searching for valves. No one seems to have questioned the functional analogy at all, so that the physiology of plant and animal nutrition was assumed rather than tested.

Academicians simultaneously escaped from and succumbed to the dangers of analogy. Although they experimented and acknowledged numerous and crucial dissimilarities, they also assumed fundamental resemblances without examining them. In matters of nutrition, analogy did substitute for experiment. Moreover, when La Hire was driven to find nonexistent valves, he let the model become axiomatic.

The Academy's circulatory analogy could be disproved by experiment and observation. Unlike Harvey's own analogies, which were didactic or were instances of similarity chosen to promote general plausibility, the analogy between sap and blood was falsified by significant dissimilarities. As Canguilhem has pointed out:

A good hypothesis is not always that which leads rapidly to its own confirmation, … It is that which obliges the researcher, by dint of an unforeseen discord between the explanation and the description, either to correct the description or to reconstruct the schema of explanation.… [I]n biology the models which have the chance of being the best are those which halt our latent tendency to identify the organic with its model[.]^[42]

The hypothesis of the circulation of sap had merit in drawing attention to a central problem in botany, namely, how sap rose at all.

It further represents an early attempt to use biology itself as a model for biological explanation. Academicians helped transform botany by finding new explanatory models. But when the biological model failed to supply a causal mechanism, academicians resorted to nonbiological causes, a fruitful reliance that affirmed the interdisciplinary character of scientific explanation. It was not "the impossibility of explaining how the vegetable machine works solely by the laws of motion"^[43] that drove botanists to zoology for inspiration. Rather an incomplete correlation between the zoological model and the vegetable explicandum forced savants back to nonbiological explanation. Seventeenth-century botanists and anatomists found that mechanics, physics, and chemistry were necessary weapons in their explanatory armory. Far from demonstrating a failure of self-image in the biological sciences, the resort to chemistry and physics exemplifies the cross-disciplinary fertilization so important for early modern science, whose practitioners were adept in many fields. Moreover, it reveals a nondogmatic use of analogy. Academicians proved some resemblances but also identified crucial dissimilarities that led them to identify an important scientific problem whose solution lay outside the original analogy. In so doing, they transformed analogical reasoning into comparative method.