Fairbanks and Rytting (2001) Controversies on Mendel Work

16
737 American Journal of Botany 88(5): 737–752. 2001. INVITED SPECIAL PAPER MENDELIAN CONTROVERSIES: A BOTANICAL AND HISTORICAL REVIEW 1 DANIEL J. FAIRBANKS 2,4 AND BRYCE RYTTING 3 2 Department of Botany and Range Science and 3 School of Music, Brigham Young University, Provo, Utah 84602 USA Gregor Mendel was a 19 th century priest and botanist who developed the fundamental laws of inheritance. The year 2000 marked a century since the rediscovery of those laws and the beginning of genetics. Although Mendel is now recognized as the founder of genetics, significant controversy ensued about his work throughout the 20 th century. In this paper, we review five of the most contentious issues by looking at the historical record through the lens of current botanical science: (1) Are Mendel’s data too good to be true? (2) Is Mendel’s description of his experiments fictitious? (3) Did Mendel articulate the laws of inheritance attributed to him? (4) Did Mendel detect but not mention linkage? (5) Did Mendel support or oppose Darwin? A synthesis of botanical and historical evidence supports our conclusions: Mendel did not fabricate his data, his description of his experiments is literal, he articulated the laws of inheritance attributed to him insofar as was possible given the information he had, he did not detect linkage, and he neither strongly supported nor opposed Darwin. Key words: Darwin; fabrication of data; fictitious experiments; laws of inheritance; linkage; Mendel. The science of genetics traces its origin to Gregor Mendel’s classic experiments with the garden pea (Pisum sativum L.) and common bean (Phaseolus vulgaris L.). Mendel presented his findings to the Bru ¨nn Natural History Society in two lec- tures in the spring of 1865 and then published the lectures in the following year as a single paper under the title ‘‘Versuche u ¨ber Pflanzen-Hybriden’’ (Experiments on Plant Hybrids), hereafter referred to as ‘‘Versuche’’ (Mendel, 1866). ‘‘Versu- che’’ contains data from eight years of experimentation, sta- tistical analysis of those data, and mathematical models of the fundamental laws of inheritance. Although his paper would eventually become the foundation for the science of genetics, Mendel did not live to see that day. He died in relative ob- scurity in 1884, 16 years before his work became widely known. This is not to say that ‘‘Versuche’’ was entirely ig- nored: it was cited at least 15 times between 1865 and 1899 (Olby, 1985). However, no one recognized its relevance to the science of inheritance until 1900 when three European bota- nists, Hugo de Vries, Carl Correns, and Erich von Tschermak, independently observed the same phenomena and arrived at the same interpretation as Mendel. In the first years of the 20 th century, genetics established itself as a core discipline in bi- ology and Mendel’s work finally began to receive widespread recognition. The year 2000 marks a century since the discovery of Men- del’s work and the birth of genetics. During that century, Men- del’s name became indispensable to science. The fundamental laws of inheritance are now known as Mendel’s laws, and the science on which they are based is called Mendelian genetics. However, because Mendel’s importance was unrecognized dur- ing his lifetime, little original information about his scientific 1 Manuscript received 11 August 2000; revision accepted 25 January 2001. The authors thank Anna Matalova ´, Director of the Mendelianum Museum Moraviae in Brno, Czech Republic for permitting examination of original copies of books and records, and W. Ralph Andersen, James F. Crow, Blair R. Holmes, Duane E. Jeffery, Alan F. Keele, L. Eugene Robertson, and Mar- cus A. Vincent for their assistance and suggestions. This research was sup- ported by funds from an Alcuin Fellowship to DJF from Brigham Young University. 4 Author for reprint requests (mail: Danielp[email protected]). work was preserved. Most unfortunately, his scientific records were apparently burned around the time of his death (Olby, 1985; Orel, 1996). In part because of the paucity of original documents, con- troversy plagued discussions of Mendel’s work throughout the 20 th century. Some authors praise Mendel as a brilliant sci- entist whose work was ahead of its time, others are critical of his methods, and a few claim he was a fraud. There is sub- stantial disagreement about his objectives, the accuracy of his presentation, the statistical validity of his data, and the rela- tionship of his work to evolutionary theories of his day. In the following pages we address five of the most contentiously de- bated issues by looking at the historical record through the lens of current botanical science: (1) Are Mendel’s data too good to be true? (2) Is Mendel’s description of his experiments fictitious? (3) Did Mendel articulate the laws of inheritance attributed to him? (4) Did Mendel detect but not mention link- age? (5) Did Mendel support or oppose Darwin? We begin with a brief overview of Mendel’s data. When quoting Mendel’s paper in English, we use Sherwood’s En- glish translation (Stern and Sherwood, 1966) as is customary for authors who write about Mendel’s work in English. We refer to the original German when necessary. A BRIEF REVIEW OF MENDEL’S DATA Mendel chose Pisum for his work after preliminary experiments with sev- eral plant species and an examination of botanical literature on plant hybrid- ization, particularly C. F. Ga ¨rtner’s (1849) Versuche und Beobachtungen u ¨ber die Bastarderzeugung im Pflanzenreiche (Experiments and Observations on Hybrid Production in the Plant Kingdom). For his hybridization experiments, Mendel selected 22 pea varieties that he had confirmed through two years of testing to be true-breeding. He reported data from hybridization experiments on seven traits that differed among the varieties. In the following list of these traits, we include some information that is not in Mendel’s paper but is per- tinent for our later discussions, such as the modern designations of the genes Mendel studied and the chromosomes on which they reside: 1. Seed shape. In mature seeds, the dominant phenotype is a smooth or slightly indented round seed, and the recessive phenotype is a wrinkled angular seed. The varieties with wrinkled angular seeds were classified

Transcript of Fairbanks and Rytting (2001) Controversies on Mendel Work

  • 737

    American Journal of Botany 88(5): 737752. 2001.INVITED SPECIAL PAPER

    MENDELIAN CONTROVERSIES: A BOTANICAL ANDHISTORICAL REVIEW1

    DANIEL J. FAIRBANKS2,4 AND BRYCE RYTTING3

    2Department of Botany and Range Science and 3School of Music, Brigham Young University, Provo, Utah 84602 USA

    Gregor Mendel was a 19th century priest and botanist who developed the fundamental laws of inheritance. The year 2000 markeda century since the rediscovery of those laws and the beginning of genetics. Although Mendel is now recognized as the founder ofgenetics, significant controversy ensued about his work throughout the 20th century. In this paper, we review five of the most contentiousissues by looking at the historical record through the lens of current botanical science: (1) Are Mendels data too good to be true? (2)Is Mendels description of his experiments fictitious? (3) Did Mendel articulate the laws of inheritance attributed to him? (4) DidMendel detect but not mention linkage? (5) Did Mendel support or oppose Darwin?

    A synthesis of botanical and historical evidence supports our conclusions: Mendel did not fabricate his data, his description of hisexperiments is literal, he articulated the laws of inheritance attributed to him insofar as was possible given the information he had, hedid not detect linkage, and he neither strongly supported nor opposed Darwin.

    Key words: Darwin; fabrication of data; fictitious experiments; laws of inheritance; linkage; Mendel.

    The science of genetics traces its origin to Gregor Mendelsclassic experiments with the garden pea (Pisum sativum L.)and common bean (Phaseolus vulgaris L.). Mendel presentedhis findings to the Brunn Natural History Society in two lec-tures in the spring of 1865 and then published the lectures inthe following year as a single paper under the title Versucheuber Pflanzen-Hybriden (Experiments on Plant Hybrids),hereafter referred to as Versuche (Mendel, 1866). Versu-che contains data from eight years of experimentation, sta-tistical analysis of those data, and mathematical models of thefundamental laws of inheritance. Although his paper wouldeventually become the foundation for the science of genetics,Mendel did not live to see that day. He died in relative ob-scurity in 1884, 16 years before his work became widelyknown. This is not to say that Versuche was entirely ig-nored: it was cited at least 15 times between 1865 and 1899(Olby, 1985). However, no one recognized its relevance to thescience of inheritance until 1900 when three European bota-nists, Hugo de Vries, Carl Correns, and Erich von Tschermak,independently observed the same phenomena and arrived atthe same interpretation as Mendel. In the first years of the 20thcentury, genetics established itself as a core discipline in bi-ology and Mendels work finally began to receive widespreadrecognition.

    The year 2000 marks a century since the discovery of Men-dels work and the birth of genetics. During that century, Men-dels name became indispensable to science. The fundamentallaws of inheritance are now known as Mendels laws, and thescience on which they are based is called Mendelian genetics.However, because Mendels importance was unrecognized dur-ing his lifetime, little original information about his scientific

    1 Manuscript received 11 August 2000; revision accepted 25 January 2001.The authors thank Anna Matalova, Director of the Mendelianum Museum

    Moraviae in Brno, Czech Republic for permitting examination of originalcopies of books and records, and W. Ralph Andersen, James F. Crow, BlairR. Holmes, Duane E. Jeffery, Alan F. Keele, L. Eugene Robertson, and Mar-cus A. Vincent for their assistance and suggestions. This research was sup-ported by funds from an Alcuin Fellowship to DJF from Brigham YoungUniversity.

    4 Author for reprint requests (mail: [email protected]).

    work was preserved. Most unfortunately, his scientific recordswere apparently burned around the time of his death (Olby,1985; Orel, 1996).

    In part because of the paucity of original documents, con-troversy plagued discussions of Mendels work throughout the20th century. Some authors praise Mendel as a brilliant sci-entist whose work was ahead of its time, others are critical ofhis methods, and a few claim he was a fraud. There is sub-stantial disagreement about his objectives, the accuracy of hispresentation, the statistical validity of his data, and the rela-tionship of his work to evolutionary theories of his day. In thefollowing pages we address five of the most contentiously de-bated issues by looking at the historical record through thelens of current botanical science: (1) Are Mendels data toogood to be true? (2) Is Mendels description of his experimentsfictitious? (3) Did Mendel articulate the laws of inheritanceattributed to him? (4) Did Mendel detect but not mention link-age? (5) Did Mendel support or oppose Darwin?

    We begin with a brief overview of Mendels data. Whenquoting Mendels paper in English, we use Sherwoods En-glish translation (Stern and Sherwood, 1966) as is customaryfor authors who write about Mendels work in English. Werefer to the original German when necessary.

    A BRIEF REVIEW OF MENDELS DATA

    Mendel chose Pisum for his work after preliminary experiments with sev-eral plant species and an examination of botanical literature on plant hybrid-ization, particularly C. F. Gartners (1849) Versuche und Beobachtungen uberdie Bastarderzeugung im Pflanzenreiche (Experiments and Observations onHybrid Production in the Plant Kingdom). For his hybridization experiments,Mendel selected 22 pea varieties that he had confirmed through two years oftesting to be true-breeding. He reported data from hybridization experimentson seven traits that differed among the varieties. In the following list of thesetraits, we include some information that is not in Mendels paper but is per-tinent for our later discussions, such as the modern designations of the genesMendel studied and the chromosomes on which they reside:

    1. Seed shape. In mature seeds, the dominant phenotype is a smooth orslightly indented round seed, and the recessive phenotype is a wrinkledangular seed. The varieties with wrinkled angular seeds were classified

  • 738 [Vol. 88AMERICAN JOURNAL OF BOTANY

    at the time as Pisum quadratum. The gene Mendel studied that governsthis trait is r on chromosome 7.

    2. Cotyledon color. In mature seeds, the dominant phenotype is yellowcotyledon color, and the recessive phenotype is green cotyledon color.The gene Mendel studied that governs this trait is i on chromosome 1.

    3. Seed coat color. The dominant phenotype is a colored-opaque seed coat,and the recessive phenotype is a colorless-transparent seed coat. Thegene Mendel studied that governs this trait is a on chromosome 1. Men-del noted that in his experiments, variation for seed coat color was al-ways associated with variations for flower color and axillary pigmen-tation. He always found colored seed coats, colored flowers, and antho-cyanin pigmentation at the axils of the stipules on the same plants, andcolorless seed coats, white flowers, with no axillary pigmentation on thesame plants in both parents and progeny. This complete association ofphenotypes that Mendel observed was a case of pleiotropy, which Men-del observed because he studied alleles of the a gene. Pea researchersin the early 1900s reported epistatic interactions of these traits in a fewexperiments with other genes (White, 1917). However, variation forseed-coat color, flower color, and axillary pigmentation in most pea va-rieties is due to variation for alleles of the a gene, and consequently thepleiotropic association of these three traits is complete in most experi-ments.

    4. Pod shape. The dominant phenotype is inflated pods, due to a parchmentlayer inside the pod, and the recessive phenotype is constricted pods,due to the absence of the parchment layer. The constricted-pod varietieswere classified at the time as Pisum saccharatum, currently known asedible-pod sugar peas. The gene Mendel studied that governs this traitis either v on chromosome 4 or p on chromosome 6.

    5. Pod color. The dominant phenotype is green unripe pods with greenvenation in the leaves, and the recessive phenotype is yellow unripepods with yellow venation in the leaves. The gene Mendel studied thatgoverns this trait is gp on chromosome 5.

    6. Flower position. The dominant phenotype is flowers borne at upper ax-illary positions along the plant, and the recessive phenotype is terminalflowers with stem fasciation near the plant apex. Varieties with terminalflowers were classified at the time as Pisum umbellatum. The gene Men-del studied that governs this trait is fa on chromosome 4.

    7. Stem length. The dominant phenotype is a long stem between the in-ternodes, causing a tall plant, and the recessive phenotype is short in-ternodes, causing a dwarf plant. Some of Mendels varieties had semi-dwarf phenotypes; however, he conducted experiments on stem lengthonly with tall and dwarf parents. The gene Mendel studied that governsthis trait is le on chromosome 4.

    The first two of these traits are considered seed traits because they areobserved in the seed cotyledons, which consist of embryonic tissue. Becauseeach seed embryo is genetically a different individual, seed-trait phenotypesmay differ among the seeds on a single heterozygous plant. The remainingfive traits are considered plant traits because they are observed in wholeplants. The seed coat consists of maternal rather than embryonic tissue, so itsphenotype is the same on every seed coat of a single plant. Thus, it is con-sidered a plant trait along with its pleiotropic counterparts, flower color andaxillary pigmentation.

    The data from Versuche are summarized in Table 1. Mendel reportedthe results of the F1 and F2 generations of seven monohybrid experiments,one for each trait, and observed near 3 : 1 phenotypic ratios in the F2 gener-ation of each experiment (Experiments 17 in Table 1). He allowed some ofthe F2 plants with the dominant phenotype to self-pollinate and obtained F3progeny from which he determined the genotypes of those F2 plants. In eachcase, the ratio of homozygotes to heterozygotes among the F2 individuals wasclose to 2 : 1 (Experiments 815 in Table 1).

    He then conducted dihybrid or trihybrid experiments for all combinationsof the traits but reported data for only one dihybrid experiment (for seed shapeand cotyledon color; Experiments 16a and 16b in Table 1) and one trihybridexperiment (for seed shape, cotyledon color, and seed coat color; Experiment17 in Table 1). In both experiments, he planted the F2 seeds and allowed the

    F2 plants to self-fertilize, then determined the genotypes of the F2 individualsfrom the phenotypes of their F3 progeny. The phenotypic and genotypic ratiosfor both the dihybrid and trihybrid experiments were consistent with the lawsof segregation and independent assortment.

    He also conducted a series of testcrosses and reported data for four dihybridtestcrosses for seed shape and cotyledon color and one dihybrid testcross forflower color and stem length (Experiments 1822 in Table 1). In the dihybridtestcrosses for seed shape and cotyledon color, he hybridized F1 plants thatwere doubly heterozygous with the homozygous parental genotypes in all fourpossible combinations of reciprocal crosses. He carried each experimentthrough the number of generations required to determine the genotypes of thetestcross progeny. In all four experiments, he observed near 1 : 1 : 1 : 1 ratiosin the testcross progeny. He also observed a near 1 : 1 : 1 : 1 ratio in theprogeny of the dihybrid testcross for flower color and stem length.

    Mendel reported general non-numeric results for several experiments withthe common bean (Phaseolus vulgaris) and stated that in most cases the re-sults were similar to those obtained with Pisum. He also reported the numericresults of one preliminary experiment for flower color in Phaseolus. He hy-bridized a variety that had colored flowers with one that had white flowers,and in the F2 generation he observed 30 plants with colored flowers and onewith white flowers. He interpreted these results as adhering to either a 15 : 1ratio for two factors, or a 63 : 1 ratio for three factors.

    Having reviewed Mendels data, we now address five of the most importantcontroversies about his work.

    ARE MENDELS DATA TOO GOOD TO BE TRUE?

    In 1902, two years after Mendels work was rediscovered,W. F. R. Weldon suspected that Mendels results were veryclose to expected values and tested this suspicion with Pear-sons newly developed x2 test. He concluded that Mendelsobserved ratios were astonishingly close to his expectations(Weldon, 1902). Weldons analysis created a brief controversyand was quickly forgotten (Mangello, 1998). The next statis-tician to question the proximity of Mendels results to expectedvalues was Ronald A. Fisher (1936) who published a now-famous paper in which he closely examined Mendels paperand reconstructed the thought process of the experiments.Fishers analysis is careful and thorough and reveals his ad-miration for Mendels work. However, his paper is best knownfor its conclusion, the same one that Weldon had arrived at 32years earlier, that Mendels results were consistently so closeto expected ratios that the validity of those results must bequestioned. Fishers work spawned a series of papers dealingwith this issue. Citations of these papers can be found in sev-eral reviews (Edwards, 1986; Piegorsch, 1986; Di Trocchio,1991; Weiling, 1991; Nissani, 1994; Orel, 1996). Unfortunate-ly, all this effort has failed to yield a definitive solution: ac-cording to Nissani (1994, p. 182), the subject remains everybit as controversial today as it was in 1936.

    Like Weldons analysis, Fishers was based on consistentlylow x2 values produced when he subjected Mendels data tox2 tests. We present the x2 values and their associated proba-bilities with the appropriate degrees of freedom for each ofMendels independent Pisum experiments in Table 2. As Ed-wards (1986) noted, about half of all independent experimentsshould yield x2 values with probabilities ,0.5 and half withprobabilities .0.5. Of Mendels 22 experiments, only twoyield x2 values with probabilities ,0.5 and six yield x2 valueswith probabilities .0.90, indicative of the bias toward expec-tation in Mendels data (Table 2).

    As Fisher (1936) and several subsequent authors (Sturte-vant, 1965; Edwards, 1986) have pointed out, Mendels dataare most suspicious where they closely approach an incorrect

  • May 2001] 739FAIRBANKS AND RYTTINGMENDELIAN CONTROVERSIES

    TABLE 1. A summary of Mendels experiments on Pisum hybrids.

    Experiments ResultsExpected

    ratio

    F2 Generation of Monohybrid Experiments1. Seed shapea2. Cotyledon colora3. Seed coat color4. Pod shape5. Pod color6. Flower position7. Stem length

    Totals

    RoundYellowColoredInflatedGreenAxialLongDominant

    54746022

    705882428651787

    14 949

    AngularGreenWhiteConstrictedYellowTerminalShortRecessive

    18502001224299152207277

    5010

    3 : 13 : 13 : 13 : 13 : 13 : 13 : 13 : 1

    F3 Progeny Tests of F2 Individuals from Monohybrid Experiments8. Seed shape9. Cotyledon color

    Totals for seed traits

    HeterozygousHeterozygousHeterozygous

    372353725

    HomozygousHomozygousHomozygous

    193166359

    2 : 12 : 12 : 1

    10. Seed coat color11. Pod shape12. Pod color13. Flower position14. Stem length15. Pod color (repeat)Totals for plant traits

    HeterozygousHeterozygousHeterozygousHeterozygousHeterozygousHeterozygousHeterozygous

    647160677265

    399

    HomozygousHomozygousHomozygousHomozygousHomozygousHomozygousHomozygous

    362940332835

    201

    2 : 12 : 12 : 12 : 12 : 12 : 12 : 1

    F2 Generation of Dihybrid Experiment for Seed Shape (A, a) and Cotyledon Color (B, b)16a. Phenotypic Round, yellow

    Round, green315108

    Angular, yellowAngular, green

    10132 9 : 3 : 3 : 1

    16b. Genotypicb ABAaBaBAbab

    3860283530

    AbbAaBbaBbAaB

    65138

    6867

    1 : 2 : 1 : 2 : 4 : 2 : 1 : 2 : 1F2 Generation of Trihybrid Experiment Seed Shape (A, a), Cotyledon Color (B, b), and Seed Coat Color (C, c)

    17. Genotypicb

    ABCABcAbCAbcaBCaBcabCabc

    814

    911

    81010

    7

    ABCcAbCcaBCcabCcABbCABbcaBbCaBbcAaBCAaBcAabCAabc

    221725201518192414182014

    ABbCc 45aBbCc 36AaBCc 38AabCc 40AaBbC 49AaBbc 48

    AaBbCc 78

    1 : 2 : 1 : 2 : 4 : 2 : 1 : 2 : 1 : 2 : 4 : 2 : 4 : 8 : 4 : 2 : 4 : 2 : 1 : 2 : 1 : 2 : 4 : 2 : 1 : 2 : 1Progeny Genotypesb of Reciprocal Dihybrid Testcross Experiments for Seed Shape (A, a) and Cotyledon Color (B, b), (Female Parent Listed First)

    18. AaBb 3 AB AB 20 ABb 23 AaB 25 AaBb 22 1 : 1 : 1 : 119. AB 3 AaBb20. AaBb 3 ab21. ab 3 AaBb

    ABAaBbAaBb

    253124

    ABbAabAab

    192625

    AaBaBbaBb

    222722

    AaBbabab

    212627

    1 : 1 : 1 : 11 : 1 : 1 : 11 : 1 : 1 : 1

    Progeny Genotypesb of Dihybrid Testcross Experiments for Flower Color (A, a) and Stem Length (B, b), (Female Parent Listed First)22. Aab 3 aBb AaBb 47 Aab 40 aBb 38 ab 41 1 : 1 : 1 : 1a Mendel also reported data for ten individual plants from each of these experiments to illustrate the variation among plants. Because these data

    are a subset of Experiments 1 and 2, we have not included the data for individual plants in this table.b Throughout this table, we have used Mendels genotypic designations, which are A 5 homozygous for A, Aa 5 heterozygous for A, and a, a 5

    homozygous for a, etc. Experiments 16a and 16b carry the same number because the data are from the same plants. The total number of individualsin Experiment 16b is less than in Experiment 16a because some F3 seeds failed to germinate and others produced plants that failed to bear seeds.

    expected ratio. After presenting the F2 segregation ratios of hisseven monohybrid experiments, Mendel proposed that two-thirds of the F2 individuals with the dominant phenotypeshould be hybrids (heterozygotes) and the remaining thirdshould be constant (homozygotes) for the trait in question,giving a ratio of 2 : 1. To test this hypothesis, he allowed F2plants with the dominant phenotypes to self-fertilize, then ob-

    served the phenotypic traits of the F3 progeny. For the seedtraits, cotyledon color and seed shape, this was a relativelyeasy task because the cotyledons of the seeds on the F2 plantsdisplayed the F3 phenotypes at maturity, so there was no needfor him to grow F3 plants to score these traits. The five planttraits presented some difficulty because the F3 phenotypescould only be scored in the F3 plants. Because of limited gar-

  • 740 [Vol. 88AMERICAN JOURNAL OF BOTANY

    TABLE 2. Results of chi-square tests for Mendels Pisum sativum ex-periments.

    Experimentfrom

    Table 1

    Degreesof

    freedom x2 Probability

    1.2.3.4.5.6.7.8.9.

    10.11.12.13.14.15.

    111111111111111

    0.13140.00750.19540.03180.22530.17480.30330.17350.42490.32000.84502.00000.00501.28000.1250

    0.71690.93100.65840.85860.63500.67590.58180.67710.51450.57160.35800.15730.94360.25790.7237

    16a.16b.17.18.19.20.21.22.

    38

    2633333

    0.47002.8110

    15.32240.57780.86210.61820.53061.0843

    0.92540.94570.95110.90150.83460.89230.91210.7809

    den space, Mendel chose 100 F2 plants with the dominant phe-notype for each of the five traits and grew ten F3 descendentsfrom each of these plants. If all ten F3 descendents had thedominant phenotype, he classified the F2 plant as constant (ho-mozygous); if the F3 descendents had both dominant and re-cessive phenotypes, he classified the F2 plant as hybrid (het-erozygous). One of the experiments (for pod color) yieldedresults that Mendel felt were too far from the predicted ratioof 2 : 1, so he repeated the experiment and obtained resultsthat were more acceptable to him. By the conclusion of thisset of six experiments (Experiments 1015 in Tables 1 and 2),Mendel had scored the progeny of 600 F2 plants, 399 classifiedas heterozygotes, and 201 classified as homozygotes, a ratiothat was extremely close to his predicted 2 : 1 ratio.

    Fisher (1936) explained that although the predicted ratio of2 : 1 is genotypically correct, Mendel should have misclassi-fied some heterozygotes as homozygotes:

    In connection with these tests of homozygosity by ex-amining ten offspring formed by self fertilization, it isdisconcerting to find that the proportion of plants mis-classified by this test is not inappreciable. If each off-spring has an independent probability, .75, of displayingthe dominant character, the probability that all ten willdo so is .7510, or 0.0563. Consequently, between 5 and6 percent of the heterozygous parents will be classifiedas homozygotes, and the expected ratio of segregatingto nonsegregating families is not 2 : 1, but 1.8874 :1.1126, or approximately 377.5 : 222.5 out of 600. Nowamong the 600 plants tested by Mendel 201 were clas-sified as homozygous and 399 as heterozygous. Althoughthese numbers agree extremely closely with his expec-tation of 200 : 400, yet, when allowance is made for thelimited size of the test progenies, the deviation is one tobe taken seriously. . . . We might suppose that sampling

    errors in this case caused a deviation in the right direc-tion, and of almost exactly the right magnitude, to com-pensate for the error in theory. A deviation as fortunateas Mendels is to be expected once in twenty-nine trials.

    Fisher (1936, pp. 125126)Later in his paper, Fisher cited these experiments once again

    as evidence that Mendels data were questionable:A serious and almost inexplicable discrepancy has, how-ever, appeared, in that in one series of results the num-bers observed agree excellently with the two to one ra-tio, which Mendel himself expected, but differ signifi-cantly from what should have been expected had histheory been corrected to allow for the small size of histest progenies. To suppose that Mendel recognized thistheoretical complication, and adjusted the frequenciessupposedly observed to allow for it, would be to contra-vene the weight of evidence supplied in detail by hispaper as a whole. Although no explanation can be ex-pected to be satisfactory, it remains a possibility, amongothers that Mendel was deceived by some assistant whoknew all too well what was expected. This possibility issupported by independent evidence that the data of most,if not all, of the experiments have been falsified so as toagree closely with Mendels expectations.

    Fisher (1936, p. 132)The last sentence (or part of it) is the most frequently quoted

    passage from Fishers paper and is often quoted out of context.Fisher did not accuse Mendel of fraud, nor did he claim thatMendels description of his experiments was fictitious, as laterhistorians were to do. Nonetheless, he suspected that an assis-tant manipulated the data and he was most disturbed by thefact that Mendels data in the F2 progeny tests were biasedtoward a 2 : 1 ratio rather than the ratio expected when thepresumed effect of Mendels misclassification is taken into ac-count.

    A x2 test of Mendels observed values of 399 and 201 andFishers expected values of 377.5 and 222.5 yields a x2 valueof 3.3020 with one degree of freedom, which is not statisti-cally significant because its probability is 0.0692. However,the probability derived from a x2 test is the probability of adeviation as great as or greater than the observed deviation ineither direction from the expected value. Fishers calculationof a probability of one in 29 that Mendel would observe thedeviation he did assumes a deviation of both the magnitudeand in the direction he observed. This halves the probabilityto 0.0346, which corresponds closely to one in 29 trials.

    Fisher (1936) stated his assumption that the probability of0.75 for each individual displaying the dominant phenotyperequired independence. Weiling (1986, 1989) argued that Men-del sampled ten seeds per plant without replacement in the F3progeny tests, and that the sampling, therefore, was not inde-pendent. He assumed that the average pea plant in Mendelsexperiments had 30 seeds per plant, 23 of which had the dom-inant phenotype (0.75 3 30 5 22.5, rounded to 23). Based onthis assumption, Weiling determined that the average proba-bility of misclassification was 23/30 3 22/29 3 21/28 3 20/27 3 19/26 3 18/25 3 17/24 3 16/23 3 15/22 3 14/21 50.0381, instead of 0.0563 as determined by Fisher.

    However, although Weilings estimate is correct for a plantwith 30 seeds, 23 of which have the dominant phenotype, it

  • May 2001] 741FAIRBANKS AND RYTTINGMENDELIAN CONTROVERSIES

    cannot be used to estimate the average probability of misclas-sification for a population of plants. For any particular numberof seeds per plant, the average probability of misclassificationmust be determined as the sum of the probabilities of mis-classification for all possible combinations weighted by theexpected frequencies of those combinations according to thebinomial distribution. When this is done, the average proba-bility of misclassification is consistently 0.0563. In otherwords, if Mendels data are from random seed samples col-lected from a binomially distributed population, Fishers esti-mate of 0.0563 as the probability of misclassification is cor-rect, even when the effect of sampling seeds without replace-ment is taken into account.

    Because Mendel provided data for each of the six experi-ments with 100 F3 plants, we can partition the x2 test to ex-amine each experiment individually, then calculate a series ofx2 values which can then be summed and the probability de-termined with six degrees of freedom. Fisher (1936) parti-tioned these experiments in his calculation of the x2 values inTable 5 of his paper, but did not partition them when he con-cluded that [a] deviation as fortunate as Mendels is to beexpected once in twenty-nine trials. As Piegorsch (1983)pointed out, when x2 values are calculated with Fishers ex-pectations after correction for misclassification and partition-ing of the six experiments, the summed x2 value is 7.6582with a corresponding probability of 0.2642 with six degreesof freedom. When this probability is halved to 0.1321, it cor-responds to a probability of about one in 7.6 trials, which isnot statistically significant and much less serious than that im-plied by Fishers estimate of one in 29 trials.

    The proximity of Mendels F3 progeny data to an incorrectexpectation is not as questionable as it might seem whenviewed in a botanical context. Fishers analysis is based onthe assumption that Mendel scored exactly ten F3 progenyfrom every F2 plant in his experiments for plant traits. How-ever, had Mendel sown exactly ten F3 seeds from each F2 plant,he would have scored fewer than ten F3 progeny in some casesbecause of losses due to germination failure, and his misclas-sification of heterozygotes as homozygotes would have beeneven greater than that proposed by Fisher. For example, inexperiments with nine plants, Mendel would have misclassi-fied on average 7.51% of the heterozygotes as homozygotes(0.759 5 0.0751). On the other hand, Mendel probably sowedmore than ten seeds in a space to be occupied by ten plants,then thinned the seedlings to ten to ensure that there were tenF3 progeny from each F2 plant. Indeed, Mendels descriptionof his method, von jeder 10 Samen angebaut is most ap-propriately translated as 10 seeds were cultivated, ratherthan 10 seeds were sown.

    Had Mendel sown more than ten seeds from each F2 plant,then he could have scored two of the plant traits in seedlingsbefore thinning. Differences in stem length, as Mendel notedin his paper, can be easily scored in seedlings a few days aftergermination. Also, as Mendel further noted in his paper, var-iation for seed-coat color was perfectly correlated with varia-tion for axillary pigmentation in his experiments. Mendelcould score F3 plants for the presence or absence of axillarypigmentation as early as two to three weeks after germinationand identify the phenotypes for flower color and seed-coatcolor that the plants would attain if grown to the floweringstage or to maturity. Indeed, almost half of the deviation fromFishers expectations in Mendels F2 progeny tests comes fromthese two experiments, in which the 136 heterozygotes re-

    ported by Mendel exceed Fishers expectation by ten plants(nine from the experiment for stem length). If the results ofthese two experiments are excluded, x2 values for Fishers ex-pectations are not statistically significant for either summeddata (x2 5 1.3795, 1 df, P 5 0.2402) or partitioned data (x25 4.0687, 4 df, P 5 0.3968).

    Fisher raised the same concern about Mendels trihybrid ex-periment (Experiment 17 in Table 1). In the trihybrid experi-ment, Mendel determined the genotypes for all three traits(seed shape, cotyledon color, and seed-coat color) of each F2individual in the F3 progeny. He could determine the seed-shape and cotyledon-color genotypes directly from the F3seeds on the F2 plants (although he had to remove at least partof the opaque seed coats from the seeds on plants with coloredseed coats to determine cotyledon color). However, to deter-mine the F2 genotypes for seed coat color he had to grow F3plants. Fisher (1936) speculated that Mendel must have grownten F3 progeny from each F2 plant with colored seed coats, asin the F3 progeny tests for seed-coat color (Experiment 10 inTables 1 and 2). However, Mendels description of his methodis vague; he simply referred to it as further investigations[Weiteren Untersuchungen]. Had Mendel scored exactly tenF3 progeny from each F2 plant with colored seed coats, thebias in his data toward an incorrect expectation is even greaterthan in the monohybrid F3 progeny tests when the summeddata are used (x2 5 4.9617, 1 df, P 5 0.0259). When halved,the probability corresponds to about one in 77 trials.

    To explain the bias in the trihybrid experiment, Orel andHartl (1994, p. 460) suggested that if Mendel had cultivated12 seeds per plant rather than 10, then x2 5 3.0, for which P. 0.05 and the insinuation of data tampering evaporates.Fisher (1936, p. 129) recognized this possibility, stating inreference to the trihybrid experiment that if we could supposethat larger progenies, say fifteen plants, were grown on thisoccasion, the greater part of the discrepancy would be re-moved. However, even using families of 10 plants the numberrequired is more than Mendel had assigned to any previousexperiment. Fishers concern in this passage is Mendels lackof garden space for growing the large number of plants re-quired for this experiment (4730 F3 plants if he grew ten F3plants from each F2 plant that had colored seed coats).

    Hennig (2000) raised an additional concern about this ex-periment, questioning why Mendel chose seed-coat color asthe third trait for analysis in a trihybrid experiment. It createdmore work for Mendel than the other traits plant traits becausefor many of the seeds he had to remove part of the seed coatsto score seed color. Also, he would know about the first twocomponents of his trihybrid cross (pea shape and color) morethan nine months before he found out about the third (Hen-nig, 2000, p. 127).

    However, we can dismiss Fishers concern about lack ofgarden space and Hennigs concern about a nine-month delayin scoring when we consider that Mendel could distinguish theplants in question for presence or absence of axillary pigmen-tation as seedlings. Because this trait can be scored in seed-lings, it is an excellent choice for the third trait in the trihybridexperiment because it creates at most a three-week delay be-tween data collection for the first two traits and the third. Gar-den space is not as critical because many seedlings can begrown in the space occupied by a single mature plant. Mendelprobably harvested the F3 seeds for this experiment in July orearly August. Had he planted the seeds in the garden soonafter harvest, he could have scored the seedlings for axillary

  • 742 [Vol. 88AMERICAN JOURNAL OF BOTANY

    pigmentation within three weeks, long before cold weather af-fected the plants. Alternatively, Mendel had a 27.5 3 4.5 mgreenhouse available to him (Orel, 1996) and he could havegrown as many as 1520 seedlings per pot for scoring in thegreenhouse during the fall and winter months. If he had al-lotted one pot for the F3 seedlings from each F2 plant, he hadmore than ample space in the greenhouse for the 473 potsrequired for this experiment.

    When these statistical and botanical aspects of Mendels F3progeny tests are considered, there is no reason for us to ques-tion his results from these experiments. However, we must stillaccount for the bias that is evident when the data for all ofthe experiments that he reported are compared as a whole (Ta-ble 2). After Fisher, numerous authors have sought reasonableexplanations for the bias in Mendels data that do not implyfraud, but most cannot withstand botanical or historical scru-tiny.

    Wright (1966) and Beadle (1967) proposed that Mendelmight have unconsciously misclassified individuals with ques-tionable phenotypes to favor his expectations. From a botan-ical point of view, this explanation can account for only anegligible degree of bias. The five plant traits display verydistinct phenotypes in multiple positions within each plant sothat each plants phenotype can be readily identified withouterror. The two seed traits, seed shape and cotyledon color,however, are potentially subject to misclassification of phe-notypes. Round seeds often have indentations that under cer-tain circumstances might lead an untrained researcher to mis-classify a round seed as being wrinkled. Also, some seedswhose genotype should confer a round-seed phenotype do notfully develop in the pod and may appear wrinkled. However,such seeds have irregular wrinkles, and Mendel used for hisseed-shape experiments varieties with seeds that have regularangular wrinkles (in Mendels words, kantig runzlig Samen).Angular wrinkled seeds have a cube-shaped phenotype and forthis reason were classified as Pisum quadratum in Mendelsday. Under these circumstances, the number of seeds thatMendel misclassified for seed-shape phentotypes was probablynegligible.

    The trait most likely to be misclassified is cotyledon colorbecause it is subject to some degree of environmental varia-tion. Green seeds may turn yellow when mature plants are leftunharvested too long in the sun. Yellow seeds may appeargreen if they are harvested before reaching full maturity. Ma-ture pea seeds of some varieties may have segments of greenand yellow coloration in the cotyledons. Mendel was awarethat misclassification of seed color was possible: in individ-ual seeds of some plants green coloration of the albumin isless developed and can be easily overlooked. However, hedismissed the possibility of misclassification of cotyledon col-or, stating in reference to this trait that with a little practicein sorting, however, mistakes are easy to avoid (Stern andSherwood, 1966, p. 12).

    Also, because of his experimental design, Mendel onlycould have misclassified cotyledon color and seed shape in alimited number of individuals. For some of the F2 individualsin his monohybrid experiments and for all individuals in hisdihybrid and trihybrid experiments, Mendel identified not onlythe phenotypes, but also the genotypes of individual seedsthrough examination of their self-fertilized progeny. This pro-cedure would have allowed him to correct any initial pheno-typic misclassifications for individuals whose genotype hadbeen determined.

    Olby (1985) and Beadle (1967) suggested that Mendelmight have stopped counting individuals when the numberswere close to the ratios he expected. However, as Campbell(1985) and Orel (1996) pointed out, Mendel explicitly deniedthis practice when he wrote near the beginning of his paper,To discover the relationships of hybrid forms to each otherand to their parental types it seems necessary to observe with-out exception all members of the series of offspring in eachgeneration (Stern and Sherwood, 1966, p. 4).

    Despite this statement Olby (1985) claimed that the data forMendels seed shape experiment indicate that he indeed didnot count all of the individuals in this experiment:

    If Mendel stopped recording his seeds before he hadexhausted the material, one would expect that his totalswould be less than that of an average crop for the pop-ulation of mother plants grown. This is so. Mendel statedthat fully ripe pods contained between 6 and 9 seeds. Ifwe take 6 as the average number, in order to make anallowance for unripe pods, then the 7,324 seeds whichMendel harvested from 253 plants would have comefrom 1046 pods, thus giving 4 to 5 pods per plant.

    Olby (1985, p. 211)Olby considered this estimate of pods per plant to be too

    small, implying that Mendel did not count all of the seeds inthis experiment. Di Trocchio (1991) raised a similar concern:

    From a calculation made by Margaret Campbell [1976],it appears that Mendel obtained an average of 28 to 37seeds per plant. . . . If the pea plant actually producedso few seeds, this vegetable would be a rarity in themarkets! Instead, we know that each plant produces onaverage more than 60 pods, and Mendel himself informsus that his plants produced pods that contained an av-erage of 69 seeds; he would therefore have obtainedat least 400500 seeds from every plant.

    Di Trocchio (1991, p. 504)If all the information provided by Mendel on numbers of

    seeds per plant is taken into account, the average number ofseeds per plant is close to 30 (16 590 seeds divided by 550plants 5 30.16 seeds per plant from the monohybrid, dihybrid,and trihybrid experiments for seed shape and cotyledon color).Is 30 seeds per plant too low for 19th century pea cultivation?Fisher (1936, p. 123) addressed this question and quoted Dr.J. Rasmussen, a pea geneticist, who wrote to Fisher: About30 good seeds per plant is, under Mendels conditions (dryclimate, early ripening, and attacks of Bruchus pisi) by nomeans a low number.

    Olbys and Di Trocchios estimates that the average numberof seeds per pod in Mendels experiments is six to nine arebased a passage in Versuche:

    In these two experiments [Experiments 1 and 2 in Table1] each pod usually yielded both kinds of seed. In well-developed pods that contained on the average, six tonine seeds, all seeds were fairly often round (Experiment1) or all yellow (Experiment 2); on the other hand, nomore than 5 angular or 5 green ones were ever observedin one pod.

    Stern and Sherwood (1966, p. 11)According to Mendel, the average number of seeds per pod

  • May 2001] 743FAIRBANKS AND RYTTINGMENDELIAN CONTROVERSIES

    was not six to nine for all pods, but rather for well-developedpods [gut ausbebildeten Hulsen]. His point in this passagewas not to give the average number of seeds per pod in all ofhis experiments, but to illustrate variation for phenotypic ratioswithin well-developed pods that have a relatively large numberof seeds. The average number of seeds per pod in all of Men-dels experiments was probably between three and four, whichresults in an average number of about seven to ten pods perplant.

    We searched for a significant body of data collected duringthe 19th century on numbers of seeds per pod, pods per plant,and seeds per plant in Pisum sativum. The earliest data wefound were from a field test of 24 garden-pea varieties con-ducted at the New York Agricultural Experiment Station inGeneva, New York during 1888 (Curtis, 1889). The averagenumber of seeds per pod was 4.47, the average number ofpods per plant 10.82, and the average number of seeds perplant 47.18, averages that are only slightly higher than thoseMendel observed. This confirms Rasmussens claim. Of the 24varieties tested, four produced averages of less than 30 seedsper plant. Therefore, although Olbys and Di Trocchios skep-ticism about the average number of seeds per plant in Mendelsdata might be valid for modern pea varieties grown under con-ditions of high fertility, the number of seeds per plant reportedby Mendel cannot be considered unreasonably low for peavarieties grown in the 19th century.

    Another explanation of the bias in Mendels data is botan-ical. Sturtevant (1965), Thoday (1966), and Weiling (1989,1991) proposed that because pollen grains are produced intetrads that consist of the four products of a meiotic event, andbecause the mature pollen grains from a tetrad may remainjuxtaposed following dehiscence, there is a possibility thatovules in a self-pollinated pea flower may be fertilized by twoor more pollen grains from the same tetrad. This model hasoften been called the urn model because it is analogous toa person sampling items without replacement from an urn. Ifsuch an event is common in pea fertilization, it could biasgenetic data away from binomial distributions and towardmean ratios. Beadle (1967) determined that such an effect isinsufficient to explain the bias in Mendels data. There is cur-rently no empirical evidence to support the urn model in Pis-um, but it is one that can be empirically tested because itshould produce a significant deviation from a binomial distri-bution for phenotypes of individual seeds from the same pod(or plants grown from those seeds). We have initiated the nec-essary experiments but do not yet have the results.

    We believe that the most likely explanation of the bias inMendels data is also the simplest. If Mendel selected for pre-sentation a subset of his experiments that best represented histheories, x2 analyses of those experiments should display abias. His paper contains multiple references to experiments forwhich he did not report numerical data, particularly di- andtrihybrid experiments. For example, he conducted dihyrid ortrihybrid experiments for all combinations of the seven char-acters he studied. However, he reported data for only one di-hybrid and one trihybrid experiment. In his words: Severalmore experiments were carried out with a smaller number ofexperimental plants in which the remaining traits were com-bined by twos or threes in hybrid fashion; all gave approxi-mately equal results (Stern and Sherwood, 1966, p. 22). Healso conducted dihybrid testcross experiments with all seventraits but reported only those for seed shape and cotyledoncolor, and flower color and stem length: Experiments [dihy-

    brid testcrosses] on a small scale were also made on the traitsof pod shape, pod color, and flower position, and the resultsobtained were in full agreement: all combinations possiblethrough union of the different traits appeared when expectedand in nearly equal numbers (Stern and Sherwood, 1966, p.29). In his second letter to Nageli (Stern and Sherwood, 1966),Mendel described a true-breeding genotype that he obtainedin 1859 (the fourth year of his experiments) from a tetrahybridexperiment for cotyledon color, seed coat color, pod shape, andstem length. He did not report the results of this or any othertetrahybrid experiment in his paper. He described in the con-cluding remarks of his paper a pentahybrid reciprocal back-cross experiment carried through several generations, but heonly reported a few selected data from this experiment. Healso reported in his paper that he conducted experiments onthe timing of flowering, peduncle length, and brownish-redpod color but he likewise did not report the data for theseexperiments.

    Mendel made it very clear that the data reported in his paperare from a subset of experiments that he conducted. In Men-dels second letter to Nageli, he referred to his paper as theunchanged reprint of the draft of the lecture mentioned; thusthe brevity of the exposition, as is essential for a public lec-ture (Stern and Sherwood, 1966, p. 61). Had he included allof his data, the paper would have been much longer. Mendelschoice to present data from a subset of his experiments createda bias that was detected only when 20th century scientists sub-jected his data to statistical analysis.

    Several authors have been quick to label Mendel as a fraudon the basis of Fishers analysis. As examples, Orel (1996)cited articles by Doyle (1968, Too many small x2s or hanky-panky in the monastery?) and Gardner (1977, Great fakesof science), and a book by Broad and Wade (1983, Betrayersof the Truth). Orel (1996, p. 207) then stated, These selectedexamples show how great scientific achievements can be dis-credited by dilettantes who claim a combination of two incom-patibles: the rigorousness of a meticulous scientist, and falsi-fication of the results. We conclude that, although the bias inMendels experiments is evident, there are reasonable statisti-cal and botanical explanations for the bias, and insufficientevidence to indicate that Mendel or anyone else falsified thedata.

    IS MENDELS DESCRIPTION OF HIS EXPERIMENTSFICTITIOUS?

    Some authors claim that although the data in Mendels papermay be accurate, his description of the experiments is ficti-tious. This assertion stems mostly from suppositions about hismonohybrid experiments. A monohybrid experiment is one inwhich two homozygous individuals that differ from each otherin only one trait are hybridized. The F1 progeny are calledmonohybrids and are heterozygous for only one of the genesunder study.

    After reporting the results of his monohybrid experimentson each of the seven traits, Mendel wrote, In the experimentsdiscussed above, plants were used which differed in only oneessential trait [wesentliches Merkmal] (Stern and Sherwood,1966, p. 17). Several authors doubt Mendels claim and arguethat he did not conduct true monohybrid experiments. The firstto do so was William Bateson, the most ardent defender ofMendelism in the first decade of the 20th century. In a footnote

  • 744 [Vol. 88AMERICAN JOURNAL OF BOTANY

    to the Royal Horticultural Societys English translation ofMendels paper, Bateson (1913) referred to Mendels claim:

    This statement of Mendels in light of present knowledgeis open to some misconception. Though his work makesit evident that such varieties may exist, it is very unlikelythat Mendel could have had seven pairs of varieties suchthat the members of each pair differed from each otherin only one considerable character (wesentliches Merk-mal).

    Bateson (1913, p. 350)In the introduction to his paper, Fisher (1936) quoted Ba-

    tesons statement that Mendels experiments might be fictitiousand proposed that a reconstruction of Mendels experimentsmight determine whether or not they were. After a detailedanalysis and proposed reconstruction of the experiments, Fish-er concluded that there can, I believe, be no doubt whateverthat his report is to be taken entirely literally, and that hisexperiments were carried out in just the way and in much theorder that they are recounted (p. 132).

    Half a century later, Corcos and Monaghan (1984) resur-rected Batesons claim and held it to be of considerable im-portance. They entitled their paper Mendel had no truemonohybrids and concluded that Mendels monohybridexperiments were performed with varieties [that differed] inseveral traits but that in each offering he concentrated his at-tention on only one (p. 499).

    Such claims that Mendels experiments were fictitious havelittle foundation when viewed from a botanical perspective.Taken in the context of Mendels paper, we must interpret hisstatement that plants were used which differed in only oneessential trait as meaning that each pair of parental varietiesused for the monohybrid experiments differed from each otherin only one of the seven traits he studied. Much of the con-fusion on this issue arises from the fact that Pisum sativum isa domesticated species and among the many cultivated varie-ties there are several different phenotypes. Batesons claim,which other authors have accepted (Corcos and Monaghan,1984; Di Trocchio, 1991; Bishop, 1996), is based on the notionthat the pea varieties available to Mendel were so highly variedthat he could not have paired his pea varieties to create sevenmonohybrid experiments.

    Contrary to this claim, the nature of variation in pea vari-eties (both old and modern) facilitates, rather than prevents,the construction of monohybrid experiments. Pea varieties fallinto three general categories: garden varieties (also calledshelling varieties), field varieties, and sugar varieties. Mostgarden varieties have white seed coats (also white flowers andno axillary pigmentation), inflated pods, green pods, and ax-illary flowers; they vary for seed shape, cotyledon color, andstem length. White seed coats are desirable for garden varietiesbecause colored seed coats can discolor the cooked peas. In-flated pods are desirable because they facilitate shelling unripepeas from the pod. Because garden varieties typically do notvary for seed-coat color, pod shape, pod color, and flower po-sition, Mendel could easily design monohybrid experimentsamong them for seed shape, cotyledon color, and stem length.Garden varieties with all possible combinations of the differingphenotypes for these three traits were readily available in Men-dels day and still are available among modern commercialgarden varieties. Field varieties were used mostly as fodder inMendels day and they typically display the dominant pheno-

    types for all seven traits. Many sugar varieties also display thedominant phenotypes for all traits except pod shape. Thus,Mendels monohybrid experiment for pod shape may have in-cluded a field variety and a sugar variety as parents. There areseveral possibilities for Mendels monohybrid experiment forseed-coat color. He could have hybridized a field variety witha garden variety that differed only for seed-coat color. Also,sugar varieties with white seed coats were available in his day,so he could have used two sugar varieties that differed onlyfor seed-coat color. Varieties with terminal flowers have al-ways been rare novelties. The most readily available terminal-flowered variety in the 19th century was called the MummyPea and it is probably the variety that Mendel used. White(1917) described two Mummy varieties, one with white flow-ers and one with colored flowers. The colored-flower varietywas the one probably available to Mendel, and this varietydiffers from most field varieties only in flower position. ThusMendel could have easily designed a monohybrid experimentfor flower position. Most pea varieties have green pods. How-ever a few garden and sugar varieties, called gold varieties,have yellow pods. Mendel could have matched any one of thegold varieties with another garden or sugar variety in a mono-hybrid experiment. A mathematical minimum of eight varie-ties is required for seven monohybrid experiments. We con-clude that Mendel could have easily designed and conductedseven monohybrid experiments with 22 varieties at his dis-posal.

    Di Trocchio (1991) accepted the argument of Bateson(1913) and Corcos and Monaghan (1984) that Mendels ex-periments were fictitious. He then took the argument a stepfurther, claiming that instead of conducting monohybrid ex-periments, Mendel must have hybridized the 22 varieties in allpossible combinations, then disaggregated the data into ficti-tious mono-, di-, and trihybrid experiments in his presentation,for the sake of simplicity. Bishop (1996) used Di Trocchios(1991) proposed reconstruction of Mendels experiments asevidence that Mendel began his experiments in 1861 and con-ducted them over a period of four years. Mendel stated inVersuche that he conducted his experiments over a periodof eight years, and he clarified the dates as 18561863 in hissecond letter to Nageli (Stern and Sherwood, 1966). Bishopargued that the dates in Mendels second letter to Nageli musthave been wrong and that Mendel wrote them because theletter was obviously a defensive response to the latters [Na-gelis] criticism (Bishop, 1996, p. 206). Bishops claim thatMendels experiments as reported in his paper were fictitiouswas part of an attempt to demonstrate that they were inspiredby his reading Darwins (1859) On the Origin of Species byMeans of Natural Selection or the Preservation of FavouredRaces in the Struggle for Life (hereafter referred to as theOrigin). Bishop surmised that Mendel was inspired by Dar-wins work in 1861 (when Mendel may have first heard ofDarwins theory of natural selection in a lecture) and that hethereafter began his experiments to counter Darwins theoryand promote the theory of special creation.

    Di Trocchios (1991) and Bishops (1996) claim that Mendelhybridized his 22 varieties in all possible combinations runscounter to the experimental design that Mendel described andthe logic on which it is based. Mendel first tested each traitindividually in monohybrid experiments, then subsequentlycombined traits in twos and threes to determine whether thepatterns of inheritance were independent. When he began hismonohybrid experiments, he probably did not know whether

  • May 2001] 745FAIRBANKS AND RYTTINGMENDELIAN CONTROVERSIES

    or not the inheritance of one trait influenced the inheritanceof another. Monohybrid experiments were essential to his ex-perimental design if he intended to study the inheritance of aparticular trait in the absence of any possible confounding in-fluences from other differing traits. Once he had determinedthat the inheritance of each trait in isolation followed the samepattern, he could then study the patterns of inheritance forcombinations of two or more traits. Taken literally, Mendelsaccount describes a well-conceived experimental design thatwould not have been difficult for him to perform.

    DID MENDEL ARTICULATE THE LAWS OFINHERITANCE ATTRIBUTED TO HIM?

    The two laws of inheritance most often attributed to Mendelare segregation and independent assortment. The law of seg-regation, stated in modern terms, is the idea that during mei-osis two alleles of a single locus, one inherited from eachparent, pair with each other, and then segregate from one an-other into the germ cells so that each germ cell carries onlyone allele of that locus. Segregation in heterozygous individ-uals produces in equal proportions two different types of gam-etes, each with one of the two alleles. The law of independentassortment, stated in modern terms, is the idea that the seg-regation of alleles of a single locus has no influence on thesegregation of alleles at another locus. The result is completelyrandom and uniform combinations of alleles of different lociin the self-fertilized progeny of dihybrid (or multihybrid) in-dividuals.

    Several authors question Mendels articulation of these laws.Olby (1985) attributed segregation of character elements (notnecessarily what we now perceive as alleles) to Mendel:

    The whole theory rests on one inference which no oneelse had the thought of making. It was simply the pre-diction of the number of different forms that would resultfrom the random fertilisation of two kinds of egg cellsby two kinds of pollen grains. Naudin had postulated thesegregation of specific essences in the formation of germcells; Mendel postulated the segregation of characterelements.

    Olby (1985, p. 101)However, although Olby attributes the laws of inheritance

    to Mendel, he also concluded that the laws of inheritancewere only of concern to him [Mendel] in so far as they boreon his analysis of the evolutionary role of hybrids, and thatMendel did not have the conception of pairs of factors orelements determining his pairs of contrasted characters(Olby, 1979, p. 67). (Monaghan and Corcos (1990, p. 268)fully rejected the idea that Mendel articulated the laws of seg-regation and independent assortment, stating that he [Men-del] did not explain his results by employing invisible partic-ulate determiners, paired or otherwise, and that the tradi-tional Mendelian laws of segregation and independent assort-ment are not given in the paper. They also concluded thatthe first Mendelian law, the law of segregation is not presentanywhere in Mendels paper. That it cannot be found has beensaid many times by quite a few writers (p. 287). Callender(1988, pp. 4142) called it the myth of Mendels Law ofSegregation; a law not to be found in either of Mendels pa-pers, nor in his scientific correspondence, nor in any statementthat can be unambiguously attributed to him. Monaghan andCorcos (1990, 1993) claimed that two of Mendels rediscov-

    erers, DeVries (1900) and Correns (1900), were the first toarticulate the law of segregation and that Thomas Hunt Mor-gan (1913) was the first to articulate the law of independentassortment.

    The claim that Mendel did not articulate or perceive the lawof segregation in his interpretation of his experiment is basedin part on the difficulty that modern readers have finding inMendels paper statements that resemble the current conceptof segregation. Genetic terms, such as allele, locus, and chro-mosome, had not been coined in Mendels day, nor was thecellular process of meiosis understood. Therefore, we mustlook for statements in Mendels paper indicating that he per-ceived paired hereditary factors that segregate from one an-other during the formation of germ cells.

    As pointed out by Olby (1985), Hartl and Orel (1992), Oreland Hartl (1994), Weiling (1994), and Fairbanks and Andersen(1999), Mendel referred to segregation of hereditary elementsseveral times in his paper. We (along with Olby, 1985) believethat the clearest statement is in the concluding remarks ofVersuche:

    One could perhaps assume that in those hybrids whoseoffspring are variable a compromise takes place betweenthe differing elements of the germinal and pollen cellgreat enough to permit the formation of a cell that be-comes the basis for the hybrid, but that this balancebetween antagonistic elements is only temporary anddoes not extend beyond the lifetime of the hybrid plant.Since no changes in its characteristics can be noticedthroughout the vegetative period, we must further con-clude that the differing elements succeed in escapingfrom the enforced association only at the stage at whichthe reproductive cells develop. In the formation of thesecells, all elements present participate in completely freeand uniform fashion, and only those that differ separatefrom each other. In this manner the production of asmany kinds of germinal and pollen cells would be pos-sible as there are combinations of potentially formativeelements.

    Stern and Sherwood (1966, p. 4243)Those who search for statements on segregation in Mendels

    paper often overlook this paragraph, probably because it isnear the end of the paper embedded in a discussion in whichMendel attempted to reconcile his observations of predictablevariation in the offspring of hybrids with those of other hy-bridists who reported that some hybrids breed true. Also, somereprints of English translations of Mendels paper omit the partof his paper that includes this paragraph (for example, seePeters, 1959). This paragraph, however, is a remarkably lucidsummary of the law of segregation. Mendels reference topotentially formative elements [bildungsfahigen Elemente]implies the existence of invisible particulate determinants ofinherited traits. We might well view the term element [Ele-ment], which Mendel used five times in this passage, as theequivalent of the modern term allele. This view is rein-forced by Hennigs (2000) observation that Mendel used theGerman term Element only ten times in his paper, all near theend of the paper in reference to the plants genotype. Mendelsreference to the enforced association [erzwungenen Verbin-dung] of differing elements indicates that he perceived thediffering elements as being paired in hybrids (heterozygotes).His statement that differing elements separate from each oth-

  • 746 [Vol. 88AMERICAN JOURNAL OF BOTANY

    Fig. 1. Chromosomal locations of the genes that Mendel studied.

    er [sich gegenseitig ausschliessen] shows a clear understand-ing of segregation that is similar to the modern view. He alsocorrectly recognized that segregation takes place only at thestage at which the reproductive cells develop (i.e., duringmeiosis).

    This paragraph, however, reveals one aspect of Mendelsperception that differs from the modern concept of segrega-tion. According to the modern concept, alleles at a single lo-cus, whether different as in heterozygotes or the same as inhomozygotes, segregate from one another during meiosis. Theabove passage suggests that Mendel perceived segregation asan anomaly restricted to hybrids (heterozygotes). He called thediffering elements antagonistic elements [widerstrebendenElemente] whose association in the hybrid is a compromise[Vermittlung], and wrote that only those elements that differseparate from one another, statements that, rephrased in mod-ern terms, suggest that only those alleles in the heterozygouscondition, and not those in the homozygous condition, arepaired and segregate from one another.

    In the paragraph that precedes the one we cited, Mendelexplained his understanding of this concept:

    When the reproductive cells are of the same kind andlike the primordial cell of the mother [i.e., a homozygouscell], development of the new individual is governed bythe same law that is valid for the mother plant. When agerminal cell is successfully combined with a dissimilarpollen cell we have to assume that a compromise takesplace between those elements of both cells that causetheir differences. The resulting mediating cell [hetero-zygous cell] becomes the basis of the hybrid organismwhose development must necessarily proceed in accordwith a law different from that of the two parental type.

    Stern and Sherwood (1966, p. 42)Mendels apparent perception of segregation as a phenom-

    enon restricted to heterozygotes sheds light on another aspectof his paper. Although Mendel represented heterozygotes witha two-letter designation (Aa), as modern geneticists usuallyrepresent them, he consistently represented homozygotes witha single letter (A or a), rather than the two letters (AA or aa)used today. For example, Mendel represented the genotypicratio of the F2 generation of a monohybrid experiment as A 12Aa 1 a, instead of AA 1 2Aa 1 aa. According to the passageabove, Mendel may have concluded that like elements (alleles)do not pair with one another and do not segregate in plantsthat are not hybrids (i.e., are not heterozygotes), and that there-fore a single letter was an accurate way to represent suchplants. Hartl and Orel (1992, p. 250) defended Mendels un-derstanding of segregation, reminding us that Mendel was notaware of chromosomes, and when the law of segregation isstated only in terms of different alleles, rather than in termsof chromosomes, Mendels view of segregation occurringonly in the heterozygotes (i.e., with different alleles) couldeasily be defended as being completely consistent even withthe modern use of the term.

    Although many authors have overlooked the passage wecited above, in which Mendel described the law of segregation,few have missed the following often-quoted statement of in-dependent assortment, which Olby (1979) called the climax ofMendels paper. This statement appears immediately afterMendels presentation of his di- and trihybrid experiments:

    In addition, several more experiments were carried out

    with a smaller number of experimental plants in whichthe remaining traits were combined by twos and threesin hybrid fashion; all gave approximately equal results.Therefore there can be no doubt that for all traits in-cluded in the experiment, this statement is valid: Theprogeny of hybrids in which several essentially differenttraits are united represent the terms of a combinationseries in which the series for each pair of differing traitsare combined. This also shows at the same time that thebehavior of each pair of differing traits in a hybrid as-sociation is independent of all other differences in thetwo parental plants.

    Stern and Sherwood (1966, p. 22)Mendels liberal use of italics in this passage (reverse italics

    in our quotation) indicates that he wished to emphasize hisconclusion of the independent inheritance of different traits.

    DID MENDEL DETECT BUT NOT MENTIONLINKAGE?

    Linkage is defined as a significant deviation from indepen-dent assortment due to proximity of genes on the same chro-mosome. In his statement on independent assortment quotedabove, Mendel concluded that for all of the traits he studiedthe behavior of each pair of differing traits in a hybrid as-sociation is independent of all other differences in the twoparental plants (Stern and Sherwood, 1966, p. 22). Scientistshave often been intrigued by the notion that Mendel studiedseven traits in a species that has a haploid number of sevenchromosomes, implying that Mendel discovered one gene oneach of the seven chromosomes, and for this reason he ob-served independent assortment. According to some accounts,had he chosen to study just one more trait he would havedetected linkage. Dunn (1965), under the assumption thatMendel studied one gene on each of the seven chromosomes,calculated the probability of doing so as 6/7 3 5/7 3 4/7 33/7 3 2/7 3 1/7 5 0.0061 (,1%), again calling Mendelsexperimental results into question.

    However, based on genetic maps of pea chromosomes, Nils-son (1951), Lamprecht (1968), Blixt (1975), and Novitski andBlixt (1979) showed that Mendel did not study one gene perchromosome. Instead, he studied two genes on chromosome 1(a and i), no genes on chromosomes 2 and 3, either two genes(fa and le) or three genes (fa, le, and v) on chromosome 4,one gene (gp) on chromosome 5, possibly one gene (p) onchromosome 6, and one gene (r) on chromosome 7 (Fig. 1).The gene in doubt is the one that governs pod shape. Recessive

  • May 2001] 747FAIRBANKS AND RYTTINGMENDELIAN CONTROVERSIES

    alleles of the v gene on chromosome 4 and the p gene onchromosome 6 both confer constricted (unparchmented) podswhen homozygous, and it is not known which of the two Men-del studied.

    The idea that two genes on the same chromosome are nec-essarily linked is a common misconception. Genes on the samechromosome are said to be syntenic but are linked only if theyare so close to one another that the frequency of crossoversbetween them is significantly less than the frequency of re-combination for independent assortment. The genetic map dis-tance at which linkage cannot be detected depends on the typeof experimental population, the number of individuals in theexperimental population, the degree of undetected doublecrossovers, and several other factors. In most testcross exper-iments (the most reliable type of linkage experiment), linkageoften cannot be distinguished from independent assortment forgenes located more than ;60 cM (centiMorgans) apart unlessresearchers use large numbers of progeny and a mapping func-tion that is appropriate for the species under study. The twogenes that Mendel studied on chromosome 1, i, which governsseed color, and a, which governs flower color, are 204 cMapart, so distant that they assort independently. The same istrue for two of the genes that Mendel studied on chromosome4; fa, which governs flower position, and le, which governsstem length, are 121 cM apart.

    Given the traits he studied, Mendel would not have detectedlinkage if he studied the p gene, which governs pod shape, onchromosome 6. He could have observed one case of linkageif he studied the v gene, which also governs pod shape. Thev gene is located 12 cM from the le gene on chromosome 4.Novitski and Blixt (1979) compared the arguments favoringthe p gene with those favoring the v gene for Mendels studiesand concluded that either scenario was possible. Because thevarieties that Mendel used are not known, the question as towhich of these two genes he studied is not likely to be re-solved.

    However, lets suppose that Mendel did study the v geneand thus had the opportunity to observe linkage. As Novitskiand Blixt (1979) pointed out, Lamprecht (1968) reported thatobserved recombination between le and v may vary from 2.6to 38.5% and that, according to Lamprecht (1941), the muta-tion rate for v may be as high as 40%. Had either or both ofthese values been on the higher side in Mendels experiments,he would not have detected linkage for the v and le genes.

    Also, Mendel conducted his experiments with F2 progeny,which are not as reliable for detection of linkage as are test-cross progeny. Although Mendel did not report data for hisexperiment with stem length and pod shape, we are fortunatethat he described such an experiment in his second letter toNageli. The experiment was a tetrahybrid experiment in whichone parental variety had green cotyledons, white seed coats,inflated pods, and short stems, and the other had yellow cot-yledons, colored seed coats, constricted pods, and long stems.Mendel obtained a true-breeding F3 plant in 1859, the fourthyear of his experiments, that had yellow cotyledons, whiteseed coats, inflated pods, and long stems. Thus, the dominantalleles for pod shape and stem length, if they were linked,were in repulsion conformation in the F1 generation and wererecombined into coupling conformation in the true-breedingdescendent that Mendel described.

    Linkage for loci that are 12 cM apart in repulsion confor-mation may escape detection in an F2 population because re-combination frequencies differ by only 5.89% from those for

    independent assortment (Fairbanks and Andersen, 1999). If, asMendel stated, he examined a small number of individuals inthis experiment and he observed at least one recombinant type(the true-breeding descendent mentioned in his letter to Na-geli), he probably did not detect linkage, if indeed the geneshe studied were linked.

    Di Trocchio (1991, p. 506) raised another question aboutlinkage and from it drew an unusual conclusion to support hiscase that Mendels experiments were fictitious: we must de-termine why Mendel did not perform other hybridization ex-periments with an eighth, ninth, or tenth character in order totest the general validity of his law. This is a particularly in-triguing question since we know that crosses with a numberof characters higher than seven would quite surely have shownlinkage. Di Trocchio concluded that Mendel did find link-age, but he discarded it as senseless in order to concentrate onthe only evident regularitynamely, 3:1 ratio. In doing so hethus chose, from among all the characters he experimented on,the famous seven non linked traits (p. 511).

    For us to evaluate such claims, it is useful to determine howMendel chose the traits he studied. Did he choose them be-cause their inheritance obeyed the laws he wished to illustrate(as Di Trocchio claimed), or did he choose them for otherreasons? One way to answer this question is to examine whichtraits other pea hybridists who had no knowledge of Mendelswork chose to study. Roberts (1929) reviewed the work ofplant hybridists who published their results before Mendel anddescribed the traits that they studied. Among the hybridistsbefore Mendel are several who studied pea. Knight (1799,1823) studied seed coat color and described the phenomenonof dominance. Goss (1824) studied cotyledon color and de-scribed dominance, as well as segregation, although in purelyqualitative terms. Seton (1824) studied stem length and coty-ledon color. Gartner (1849) reviewed Knights work with seedcoat color, and described his own work in Pisum with stemlength, flower color, cotyledon color, and seed shape. Mendelstudied Gartners (1849) book on plant hybridization in detailbefore and during his experiments, as he indicated in his firstletter to Nageli and as evidenced by his 17 references to it inVersuche. Thus, he was familiar with both Knights andGartners pea hybridization experiments. We examined Men-dels copy of Gartners book and found numerous marginaliathroughout it. On the page facing the back cover are Mendelshandwritten notes about traits in Pisum, which, as translatedby Olby (1985) read:

    Pisum arvense: flowers solitary, wings red.Pisum arvense et sativum: pods almost cylindrical, inPisum umbellatum Mill. [terminal-flowered varieties]cylindrical and straight; in saccharatum Host. [sugar va-rieties] Straight, ensiform, constricted on both sides.(var. flexuosum Willd. sickle-shaped, seeds small, an-gular); in Pisum quadratum Mill. [wrinkled-angularseeded varieties] straight, ensiform, not constricted,seeds pressed tightly together. In Pisum sativum and ar-vense the bases of the stipules rounded and denticulate-crenate, stipules cordate. In saccharatum and quadra-tum, stipules obliquely incised, pods pressed flat. In sa-tivium, saccharatum and umbellatum, seeds round.

    Olby (1985, pp. 212213)Mendel studied all four traits that previous hybridists had

    studied, seed shape, cotyledon color, stem length, and seed-

  • 748 [Vol. 88AMERICAN JOURNAL OF BOTANY

    coat color (also flower color). His handwritten note mentionsflower color, pod shape, and seed shape, and flower positionis implied in the note by his mentioning P. umbellatum.

    The following comment from Mendels paper about themonohybrid experiments reveals information about the orderof his monohybrid experiments: Experiments 1 and 2 [seedshape and cotyledon color] have by now been carried throughsix generations, 3 and 7 [seed coat color and stem length]through five, and 4, 5, and 6 [pod shape, pod color, and flowerposition] through four (Stern and Sherwood, 1966, pp. 1516). The generation for seed traits can be scored one growingseason earlier than that for plant traits, so Mendel must haveinitiated the monohybrid experiments for seed shape, cotyle-don color, seed coat color, and stem length (the same traitsthat Gartner studied) in the first year of his hybridization ex-periments, and the monohybrid experiments for the three re-maining plant traits in the following year.

    The four traits studied by previous hybridists were the samefour traits that Mendel used in his first year of hybridization.The remaining three traits, pod shape, pod color, and flowerposition, were not among those studied by previous hybridists,and Mendel initiated experiments on them during the follow-ing year. This reconstruction argues against Di Trocchios(1991, p. 508) assertion: It is likely, in fact, that he [Mendel]planned his hybridization experiments following the checker-board method. A checkerboard of 22 3 22 squares representsall of the crosses. Several other items in Mendels paper alsoargue against Di Trocchios assertion. According to statementsin Mendels paper, the varieties he used as parents for hismonohybrid experiments for seed color did not include vari-eties with colored seed coats (because the opaque colored seedcoats prevent observation of cotyledon color), and those heused for his monohybrid experiments for stem length did notinclude those with intermediate stem lengths.

    The type of analysis that Mendel conducted relies on traitsthat display discontinuous variation, as does detection of link-age. The traits that Mendel chose to study are among only afew that varied in a discontinuous fashion among commer-cially available 19th century varieties. Mendels varieties cer-tainly differed in more than the seven traits on which he re-ported data. However, most of the other traits display contin-uous variation and are governed by multiple genes and envi-ronmental influences and cannot be easily analyzed in a simpleMendelian fashion. Mendel listed a number of these traits inhis paper, then stated that he could not clearly analyze suchtraits: However, some of the traits listed do not permit adefinite and sharp separation, since the difference rests on amore or less which is often difficult to define. Such traitswere not usable for individual experiments; these had to belimited to characteristics which stand out clearly and decisive-ly in the plants (Stern and Sherwood, 1966, pp. 56).

    Mendels choice of traits apparently was based first on thosestudied by his predecessors and second on those that had dis-tinct discontinuous phenotypic differences that permitted con-clusive analysis. Because of the locations of the genes thatgoverned such traits and the design of his experiments, it isunlikely that he could have detected linkage. There is no bo-tanical or historical evidence to support the claim that Mendelobserved and then disregarded linkage.

    DID MENDEL SUPPORT OR OPPOSE DARWIN?We addressed the four previous controversies in botanical

    contexts. This final controversy is purely historical, but it is

    widely debated by the same authors who address botanicalissues and they have related their conclusions on botanicalissues to this issue. Thus, we also are compelled to include it.Mendel and Darwin were contemporaries and both addressedevolutionary questions in their work. Twice in VersucheMendel used the term Entwicklungsgeschichte, which in theEnglish translations of Versuche is rendered as evolutionor evolutionary history. One of the passages with this termappears near the beginning of Versuche where Mendel re-ferred to his experiments as the one correct way of finallyreaching the solution to a question whose significance for theevolutionary history [Entwicklungs-Geschichte] of organicforms must not be underestimated (Stern and Sherwood,1966, p. 2).

    Thus, Mendel was clearly interested in evolution and heconsidered his experiments as relevant to an understanding ofevolution. At the time Mendel wrote his paper, the Origin waswell known and evolution through natural selection was a pop-ular topic for discussion in scientific societies. Alexander Ma-kowsky, who was a student at the same school as Mendel andwas among his closest friends, presented a lecture that favor-ably treated Darwins theory of natural selection to the BrunnNatural History Society the month before Mendel presentedthe first of two installments of his article to the Society (Ma-kowsky, 1866).

    Although there is no evidence that Darwin knew of Men-dels work, there is ample evidence that Mendel read some ofDarwins writings and that those writings may have influencedhis work. Efforts to elucidate the MendelDarwin connectionhave been underway for nearly a century (for examples, seeBateson, 1913; Iltis, 1924; Fisher, 1936; Olby, 1985; Callen-der, 1988; Bishop, 1996; Orel, 1996).

    In spite of much research, there is no consensus about Men-dels views on Darwinism. The numerous articles and com-mentaries on this topic begin with a comment in a letter toWilliam Bateson written in 1902 by Mendels nephew, Fer-dinand Schindler, who stated, He [Mendel] read with greatinterest Darwins work in German translation, and admired hisgenius, though he did not agree with all of the principles ofthis immortal natural philosopher (Orel, 1996, p. 188). Ba-teson (1913, p. 329) wrote, With the views of Darwin whichat that time were coming into prominence Mendel did not findhimself in full agreement. Fisher (1936, p. 118) believed thatMendel understood his laws to form a necessary basis forthe understanding of the evolutionary process and that hadhe [Mendel] considered that his results were in any degreeantagonistic to the theory of selection it would have been easyfor him to say this also. Sapp (1990) determined that Fishers1936 paper was the turning point for the modern synthesisof Mendelism and Darwinism. In the mid-1960s, when a largenumber of articles were published at the centennial of Men-dels paper, most authors viewed Mendel as a supporter ofDarwinism. By contrast, Olby (1979, 1985) studied the his-torical context of evolutionary thought during Mendels dayand determined that Darwins views on the role of hybrid-ization in evolution were very far removed from Mendels(Olby, 1979, p. 67). Callender (1988) and Bishop (1996) ex-pressed the most extreme views of Darwins influence on Men-del. Callender (p. 72) claimed that Mendelism came into be-ing historically as a sophisticated form of the doctrine of Spe-cial Creation and that it stood in open conflict with theDarwinian conception of evolution as descent with modifica-tion by means of Natural Selection. Bishop (p. 212) proposed

  • May 2001] 749FAIRBANKS AND RYTTINGMENDELIAN CONTROVERSIES

    that Mendels sole objective in writing his Pisum paper, pub-lished in 1866, was to contribute to the evolution controversythat had been raging since the publication of Darwins theOrigin of Species in 1859, and that Mendel was in favor ofthe orthodox doctrine of special creation. After a detailedreview of the literature on Mendels perception of evolution,Orel (1996, p. 198) determined that Mendel came across Dar-wins theory as his Pisum experiments were drawing to aclose. From his notes and from indirect evidence one can sup-pose that he did not see any conflict between this theory andhis own.

    The extreme disagreement among scholars about Mendelsview of Darwins writings is probably because Mendel wrotevery little about Darwin, and thus most claims are suppositionsabout what Mendel must have thought about Darwin. In hissurviving writings, Mendels overtly referred to Darwin onlyfour times, all in 1870, four years after the publication ofVersuche. One reference is in Mendels (1870) Hieraciumpaper and three are in his eighth and ninth letters to Nageli(Stern and Sherwood, 1966). All four references are brief andreveal neither strong support of nor opposition to Darwinstheories.

    Because Mendels major contribution to the science of ge-netics was Versuche, we will focus on how Darwin mayhave influenced Mendel before the 1866 publication of Ver-suche. Of Darwins writings, only the Origin was availableto him before 1866. Several authors have noted that Mendelspersonal copy of the Origin contains marginalia (Iltis, 1924;Moore, 1963; Voipio, 1987; Hartl and Orel, 1992; Bishop,1996; Orel, 1996). Mendel purchased a copy of the secondGerman edition, published in 1863, which was translated fromthe third English edition (Darwin, 1861). This copy containsMendels marginalia and is in the collection of the Mende-lianum Museum Moraviae in Brno.

    When Mendel began his classic experiments with pea in1856, none of Darwins works were available for him to read.According to Orel (1971, 1996), Mendel probably first heardof Darwin in September 1861 during a lecture. He might haveread the Origin during the latter part of 1862 or early 1863when the Brunn Natural History Society acquired a copy ofthe German translation of the first English edition (Darwin,1859). The 1863 publication date of Mendels personal copyof the Origin coincides with the last year of his experimentswith peas. Therefore, the Origin had no effect on the designor conduct of those experiments, although it may have influ-enced Mendels interpretation of those experiments in Ver-suche. Orel (1996), de Beer (1964), Fisher (1936), and Ba-teson (1913) concluded that Darwins influence on Mendel,primarily from the Origin, is evident in Versuche. Our com-parison of Mendels marginalia in the Origin with passages inVersuche supports this view.

    Mendels complete marginalia in the Origin have not beenpublished, although Orel (1996) discussed a few of them. Theappendix to this paper contains full German and English textsof the complete marginalia along with our commentary andcan be accessed electronically at http://ajbsupp.botany.org/v88/fairbanks.html. When quoting from the English version of theOrigin in both this manuscript and the appendix we use thethird edition text (Darwin, 1861) because this is the Englishedition from which Mendels German copy was translated.

    The marginalia consist of passages marked in pencil andtwo very brief notes in script. The marks are either single ordouble vertical lines in the margins next to passages that Men-

    del apparently found interesting. Mendel marked passages ononly 18 pages. The marked passages are clustered into twogroups. Eight of them are in Chapters 14 (five in Chapter 2Variation Under Nature), and ten of them are in Chapters8 and 9 (eight in Chapter 8 Hybridism).

    The marginalia include only two notes in script that can beattributed to Mendel. One of them is the series of numbers16713164852576263767880 written inside the backcover of the book. These may be page numbers, but our ex-amination of the corresponding pages suggests that the num-bers do not refer to pages in the Origin (see the Appendix).The other note in script is on page 1 and reads pag 302.The term pag probably is an abbreviation of the Latin wordpagina for page. On page 302 is a passage that Mendel markedwith double lines. In the original English it reads: The slightdegree of variability in hybrids from the first cross or in thefirst generation, in contrast with their extreme variability inthe succeeding generations, is a curious fact and deserves at-tention (Darwin, 1861, p. 296).

    Apparently, Mendel found this to be the most interesting ofthe passages he marked. It is the only passage he cited by pagenumber and is one of only two passages marked with doublelines. Mendels observations of uniformity in the F1 generationand predictable variability in the F2 generation and his theo-retical explanations for these phenomena form one of the keypoints of his paper. According to Orel (1996, p. 193), HereMendel must have felt some gratification in the thought thathis theory was soon to explain this curious fact.

    Of interest is Darwins explanation, which immediately fol-lows this passage and is partially included within Mendelsmark. Darwins explanation of the uniformity of hybrids in theF1 generation and the variability of their F2 offspring differssubstantially from Mendels in that Darwin places the causeon altered reproductive systems rather than constant inheritedtraits:

    For it bears on and corroborates the view which I havetaken on the cause of ordinary variability; namely, thatit is due to the reproductive system being eminently sen-sitive to any change in the conditions of life, being thusoften rendered either impotent or at least incapable ofits proper function of producing offspring identical withthe parent-form. Now hybrids in the first generation aredescended from species (excluding those long cultivated)which have not had their reproductive systems in anyway affected, and they are not variable; but hybridsthemselves have their reproductive systems seriously af-fected, and their descendants are highly variable.

    Darwin (1861, p. 296)This is one of many passages in the Origin in which Darwin

    uses the phrase conditions of life, which in Mendels Ger-man edition of the Origin is translated as Lebens-Bedingun-gun. Mendel marked three passages in the Origin with thisphrase (pages 17, 295, and 302 in his German edition; see theAppendix). The most important is the first marked passage inMendels copy of the Origin:

    It seems pretty clear that organic beings must be ex-posed during several generations to the new conditionsof life to cause any appreciable amount of variation; andthat when the organization has once begun to vary, itgenerally continues to vary for many generations.

    Darwin (1861, p. 7)

  • 750 [Vol. 88AMERICAN JOURNAL OF BOTANY

    This passage appears in the opening remarks of Chapter 1,Variation Under Domestication, in which Darwin (1861, p.7) suggested that the higher degree of variations in domesti-cated species compared to their wild counterparts is due to thedomestic productions having been raised under conditions oflife not so uniform as, and somewhat different from, those towhich the parent-species have been exposed under nature.Mendels view of this subject differed from Darwins. FromVersuche:

    Granted willingly that cultivation favors the formationof new varieties and that by the hand of man many analteration has been preserved which would have per-ished in nature, but nothing justifies the assumption thatthe tendency to form varieties is so extraordinarily in-creased that species soon lose all stability and theirprogeny diverge into an infinite number of variableforms. If the change in living conditions [Lebensbedin-gungun] were the sole cause of variability one wouldexpect that those cultivated plants that have been grownthrough centuries under almost identical conditionsshould have regained stability. This is known not to bethe case, for it is precisely among them that not only themost different but also the most variable forms arefound.

    Stern and Sherwood (1966, p. 37)Commenting on this passage, Fisher (1936, p. 134) wrote:

    The reflection of Darwins thought is unmistakable, andMendels comment is extremely pertinent, though it seems tohave been overlooked. He may at this time have read the Or-igin, but the point under discussion may equally have reachedhis notice at second hand. Indeed, Mendel probably had readthe Origin at the time he presented Versuche. This passagefrom Versuche seems to be a direct response to the passagehe marked on page 17 of the Origin. In it Mendel contradictedDarwins claim that


Cypress fairbanks isd

Cypress fairbanks isd

Fairbanks Health & Wellness

Fairbanks Health & Wellness

FAIRBANKS PLAZA - LoopNet

FAIRBANKS PLAZA - LoopNet

Controversies in Stroke Care: Acute Interventions · Controversies in Stroke Care - Acute Interventions . ... controversial topics enables an effective debate about ... Controversies

Controversies in Stroke Care: Acute Interventions · Controversies in Stroke Care - Acute Interventions . ... controversial topics enables an effective debate about ... Controversies

Public Art Controversies

Public Art Controversies

Leah Fairbanks

Leah Fairbanks

Museum Controversies

Museum Controversies

Manual Fairbanks

Manual Fairbanks

Fairbanks Manual - Weldmart

Fairbanks Manual - Weldmart

Natassja fairbanks a2_blogppt_

Natassja fairbanks a2_blogppt_

Controversies in Iui

Controversies in Iui

Controversies in Cognition

Controversies in Cognition

Slide 1 Linear and Nonlinear Device Measurements Doug Rytting.

Slide 1 Linear and Nonlinear Device Measurements Doug Rytting.

Thesis by Fairbanks

Thesis by Fairbanks

Prsa fairbanks

Prsa fairbanks

The Haines-Fairbanks Pipeline - ExploreNorthexplorenorth.com › library › military › Haines_Pipeline.pdf · The Haines-Fairbanks Pipeline The Haines-Fairbanks Pipeline Fairbanks

The Haines-Fairbanks Pipeline - ExploreNorthexplorenorth.com › library › military › Haines_Pipeline.pdf · The Haines-Fairbanks Pipeline The Haines-Fairbanks Pipeline Fairbanks

Network Analyzer Error Models and Calibration Methods by Doug Rytting

Network Analyzer Error Models and Calibration Methods by Doug Rytting

467 FAIRBANKS MORSE.pdf

467 FAIRBANKS MORSE.pdf

1 TSD-160 Introduction to Network Analyzers and Error Correction Doug Rytting 4804 Westminster Place Santa Rosa, CA 95405 707-539-1631 rytting@sbcglobal.net.

1 TSD-160 Introduction to Network Analyzers and Error Correction Doug Rytting 4804 Westminster Place Santa Rosa, CA 95405 707-539-1631 [email protected].

Group Tests and Controversies in Ability Testing Test Bias and Other Controversies.

Group Tests and Controversies in Ability Testing Test Bias and Other Controversies.