Historical perspective on developmental concepts and terminology


  • John M. Opitz,

    Corresponding author
    1. Pediatrics (Medical Genetics), Pathology, Human Genetics, Obstetrics and Gynecology, University of Utah School of Medicine, Salt Lake City, Utah
    • Correspondence to:

      J.M. Opitz, c/o Am J Med Genet, University of Utah, 419 Wakara Way, Ste. 213, Salt Lake City, UT 84108.

      E-mail: john.opitz@hsc.utah.edu

    Search for more papers by this author
  • Giovanni Neri

    1. Istituto di Genetica Medica, Università Cattolica del Sacro Cuore, Roma, Italy
    Search for more papers by this author


In their ontogeny and phylogeny all living beings are historical entities. The revolution in biology of the 18th and 19th centuries that did away with the scala naturae according to which we humans, the acme of creation, “made a little lower than the angels,” also led to the gradual realization that a humble one-celled protist (“protoctist”), such as Entamoeba histolytica of ill repute [Margulis and Chapman, 2010] has the same 4-billion-year phylogeny as that of Homo sapiens, vivid testimony to common ancestry and the relatedness of all living beings on earth. The group of medical geneticists who assembled at the NIH, Bethesda, MD this January to address terms pertaining to human ontogeny, did so in the long tradition of Sydenham, Linnaeus, Meckel, Geoffroy St-Hilaire père et fils, Wilhelm His and so many others before who had over the previous two centuries wrestled as earnestly as they could with concepts of “classification” and nomenclature of developmental anomalies. The prior massive need for classification per se in medical morphology has diminished over the years in favor of ever more sophisticated understanding of pathogenesis and cause through experimental biology and genetics; however, in the winter of 2013 it was still found prudent to respect terminological precedent on general terms while recognizing recent advances in developmental pathology requiring clarification and definition of special terms. Efforts along similar lines instigated by the German Society of Anatomists at their first meeting in Leipzig in 1887 culminated, after intense years of work by hundreds of experts and consultants under the goad of Wilhelm His, in the Basel Nomina Anatomica [BNA, His (1895)]. His, himself, stated prefatorily that the BNA had no legislative weight, only an evanescent consensus of many to be amended in the future as needed and indicated. Without hubris, no one before or after will do the same. The more substantial the consensus the more permanent the structure. After some 120 years the BNA is alive and flourishing. Now retitled Terminologia Anatomica, it has been amended and added to many times, is still in Latin but now with synonyms in English, the new lingua franca of science, for every anatomical, histological and embryological term. May our successors be equally effective. © 2013 Wiley Periodicals, Inc.


Recently (1/26–1/29/2013) a conference was convened at the United States Institutes of Health, Bethesda, hosted most graciously by Dr. Leslie G. Biesecker [see Hennekam et al., 2013, this issue]. The purpose of the meeting was to rethink nomenclature pertaining to “Birth Defects” (assigned title) last defined in 1986 at the memorable International Conference of Human Genetics in Berlin, Arno G. Motulsky President [Opitz et al., 1987]. Of course we have made “progress” since 1986; however, the more urgent the need to anticipate the future, the greater the obligation to understand the past (“you can't know where you are going if you don't know where you came from”—attributed to many in one form or another). What is all life but history? Our ontogeny from conception to death may extend to only a century including embryogenesis, organogenesis, fetal life, birth, “growth and development” (as the pediatricians are wont to say), maturity and death. But, our phylogeny is truly awe-inspiring beginning as it did from inorganic matter some 4 billion years ago, leading to the emergence of primates some 65 million years ago and the separation of chimps and humans only 4.5–5 million years ago [these two species having a difference in non-coding DNA of only 1.2%, Roberts, 2011]. Shakespeare said it most succinctly: “What's past is prologue” (The Tempest), and Santayana (1863–1952): “Those who cannot remember the past are condemned to fulfill it,” paraphrasing so many before him. What an enormous intellectual satisfaction for those of us who decades ago “did” human genetics primarily in (pediatric) genetics clinics, with dozens of unresolved “familial cases” still in our files, to be handed on a golden platter, so to speak, the resolution to one of them, the segregating, highly conserved gene involved, functionally present not only in eukaryotes but also in bacteria and archaebacteria, hence present in the last universal common ancestor (LUCA) [Opitz, 2012], hence billions of years old.

We all leave as we came, “destined” to develop into the unique human beings we were, at the end returning to the biosphere our constituent molecules, fashioned long ago by trial and error into the successful essentiality that unites us with all organized beings on earth from beginning till now.

And yet, one of the authors' (J.M.O.) maternal great-grandfather of the 19th century knew folks anciently involved in Napoleon's wars, and his teacher Emil Witschi was a student of Richard Hertwig a scion of the school of Johannes Müller (born in 1801). We truly, as dwarves, “stand on the shoulders of giants” (Newton paraphrasing, again, predecessors).

In what follows we shall attempt to review several of the issues discussed at the Bethesda meeting from a historical perspective.


They come, go, or stay. Regardless of how old, those with staying power usually are amended many times, reinterpreted, added to, corrected or rid of obsolescences. Their flux depends on many personal, cultural, sociological and scientific factors, events, discoveries, and epistemological perspectives.

Mendeleev's Periodic Table of Elements [1869] remains, now stronger than ever, representing a triumph of logic and consistency with scientifically testable properties and predictability confirmed and added to by countless hordes of chemists in many lands, with partly Latin terminology and nomenclature acceptable to all. Thus, Pl: lead from L. plumbum, hence, plumber; Aur: gold from L. aurum; Arg: silver from L. argentum….

The Linnaean system seems equally robust beginning as early as 1732 with Carl v. Linné's sexual classification of plants in his Flora Lapponica… emerging as it were “fully armored from the head of Zeus” as the Systema Naturae with the same 2013 “license plates” as in 1735 [Blunt, 2001], a remarkable accomplishment by someone who considered himself quite entitled to do so: “Deus creavit, Linnaeus disposuit.” Impeccable Latin, but the implication being: “God created (this mess, and I,) Linnaeus, (came to straighten it out).” By the time the English version of 1806 in 7 large volumes reached J.M.O.'s library (via Marcus Jacobsen) it was entitled: “A General System of Nature, through the Three Grand Kingdoms of Animals, Vegetables, and Minerals, systematically divided into their several classes, orders, genera, species and varieties with their Habitations, Manners, Economy, Structure, and Peculiarities. By Sir Charles Linné… Dictionary, explanatory of the Terms which occur in the several Departments of Natural History, by William Turton, M.D. Fellow of the Linnaean Society…” (Fig. 1A–C) The Linnaean rules, adhered to by all botanists and zoologists come with a formidable set of instructions, rules as to priority, nomenclature, manner of prototype description, a compulsory “binomial” (literally!) naming system for genus and species initially based on Latin (and Greek) but from the beginning making Linnea, Fuchsia, Gardenia, Darwinea, Banksia, Rudbeckia… acceptable.

Figure 1.

A: Linnaeus; frontispiece in the Turton English translation of 1806, no copyright. B: Characteristics of birds, Turton translation of Linnaeus, Vol. 1, 1806, no copyright (C) characteristics of fish, Turton translation of Linnaeus, Vol. 1, 1806, no copyright.

At the first meeting of the German Society of Anatomists in Leipzig in 1887 it was decided to take on the entire anatomical nomenclature, then in a considerable state of chaos. Especially in Germany which had for some decades translated traditional Latin terms into the vernacular – so: processus coracoideus (coracoid process) as Rabenschnabelfortsatz, os sacrum as Heiligbein, clavicle as Schlüsselbein…. The individual effort by Jakob Henle to introduce new, objective terms (we owe him sagittal, frontal, medial, lateral, etc.) was considered by Wilhelm His (Fig. 2) as divisive since many authors continued to use the old, others the new, yet others both, or used Henle's precedent to coin their own. At the outset, this effort required the good will of all participants, cohesion of the membership of the Society, establishment of a Commission with their elected editor who would devote years to this effort and make it his Lebensaufgabe (life's mission), and much money (10,000—DM, an enormous sum, in gold, at the time, 80% coming from various academies, 20% from the membership). While the Society had been created in Berlin, it was truly international with 129 non-German and 145 German members. Editor Krause used Gegenbaur's text [1885] as basis but pulled all additional terms that had been used by Henle, Hyrtl, Krause, and Langer; English ones from Quain and those in French from Sappey. British, French and Italian members were added to the commission and were charged to come up with terms that were appropriate (zweckmässig), succinct, precise, and characteristic of the part, unambiguous and not simply fashionable (Mode…bemächtigt), but should lead to a logical cohesion of the entire system. The evanescent nature of the terms with advancing knowledge meant the work of the commission was ongoing. Also, that a resulting nomenclature was only a recommendation, nothing to be forced. Wilhelm His regretted the loss of eponyms since they had mnemonic usefulness. Krause sent out 20 thick fascicles with some 30,000 terms and names for discussion about one half of which were finally adopted and became the renowned Basel Nomina Anatomica [His, 1895]. It is alive and strong, but in 1998 was renamed Terminologia Anatomica (et Histologica for human cytology and histology).

Figure 2.

Wilhelm His the Elder (1831–1904).

Closer to home is the ISCN (International System for Human Cytogenetic Nomenclature, 2013) edited by Lisa G. Shaffer of Signature Genomics, Spokane, Jean McGowan-Jordan of Eastern Ontario and Michael Schmid of Würzburg. It has an excellent historical introduction. As a member of the American Society of Human Genetics, the Patau report of 1960 was known to one of us (J.M.O.) upon arrival in Madison in 1961 and to begin working with Drs. Klaus and Eeva Patau and David Smith. Klaus Patau's system of grouping the human chromosomes A–G became the basis of the “Proposed Standard System of Nomenclature of Human Mitotic Chromosomes”—“The Denver Conference.” In 1984 David Harnden commented that the Denver Nomenclature formed the basis for all subsequent nomenclature reports and “…remained virtually unaltered, despite the rapid developments of the last 25 years. It is fair to say that the participants at Denver did their job so well that this report has formed the cornerstone of human cytogenetic nomenclature since 1960, and the foresight and cooperation shown by these investigators have prevented much of the nomenclature confusion which has marked other areas of human genetics” (meaning the clinicians; ital. added) [Harnden in Shaffer et al., 2013] (Fig. 3).

Figure 3.

Human cytogenetic nomenclature, Paris Conference 1971. White: Negative or pale staining Q and G bands, positive R bands. Black: Positive Q and G bands, negative R bands; crosshatched: variable bands. Birth Defects Original Article Series VIII (7), Hamerton JL, Jacobs PA, Klinger HP, editors. 1973. Reprinted with the gracious permission of the National Foundation—March of Dimes.

One of the authors (J.M.O.) was present at the III International Congress of Human Genetics in Chicago in 1966 when the Nomenclature was revised and Lionel Penrose said: “It is easy to be carried away by the detectable peculiarities and to forget that much underlying variability is still hidden from view until some new technical device discloses the finer structure of chromosomes, as in the Drosophila salivary gland cells” [Harnden, 1984]. The IV International Congress of Human Genetics [1971, Paris] appointed a Standing Committee which rotates but has been continuously active over the last 40 years, proposing the “Paris Conference” of 1971, “a highly significant document in the annals of human cytogenetics” [Harnden, 1984]. The historical section 1985–1995 was written by Evans and Jacobs; while its practice sanctioned the truly perverse use of the term, for example, chromosome 14q32.2, as if humans had as many chromosomes as identifiable bands (like pigeon microchromosomes?), it did adopt Felix Mitelman's extremely useful “Guidelines in Cancer Cytogenetics.”

Syndrome Qua Pleiotropy

Syndrome: Concurrence of manifestations “characterizing a specific disease” [Webster: Woolf, 1981]. Even though rather generic, this dictionary definition does mention “a specific disease,” allowing the inference that possible chance concurrence of manifestations in one individual is excluded, “specific disease,” therefore, referring to causal entities identified epistemologically on the basis of:

  • Greater-than-chance concurrence of identical or very similar sets of manifestations in two or more individuals suggesting, at the very least, similar or identical pathogenesis, subject to causal verification through the discovery of physical, infectious, toxicological, or genetic factors.
  • “Genetic means”: Incest or other close consanguinity as presumed cause of a specific set of anomalies in an infant mostly due to homozygosity of one autosomal recessive gene or several such genes. Or, horizontal or vertical segregation of the same complex of manifestations in a family with or without parental consanguinity. Or, segregation of an X-linked condition with or without carrier manifestations. Biochemical and sophisticated molecular/cytogenetic means may now allow causal conclusions in a single (“sporadic”) case. Epigenetic modification of gene expression may give rise to syndromes. Causally complex mechanisms as in maternal diabetes may affect embryo or fetus “syndromally,” for example, with macrosomia, microtia, neural tube defect or congenital heart defect, caudal dysgenesis to the point of sirenomelia, anal atresia, radial deficiencies, etc. [Castori, 2012]. This author's (Castori's) paper will be published in a journal entitled “Molecular Syndromology” suggesting it deals largely with causally defined syndromes. In recent correspondence, Spranger [2013] who, over decades, has had an authoritative voice in these matters, quotes Sennert to the effect: Cum scire sit semper causam cognoscere, “to know a thing one has to know its cause” (semper, always).

Defining syndromes on the basis of a causal nexus is the essential means of advancing knowledge in medical/human genetics.

The late Victor McKusick liked to invoke Faber's [1923] history of the conceptual change in understanding disease entities sui generis since Thomas Sydenham (1624–1689) [Faber, 1923].

It was Sydenham's conviction that diseases be treated as objectively as species (e.g., of plants in the manner of Linnaeus), “since morbid phenomena are subject to laws.” The living conditions in Sydenham's days were so appalling that it gave him more than adequate opportunity to delineate acute diseases such as the plague, smallpox, cholera, phthisis, and dysentery. His confrontation with chronic diseases was far more problematic (too little time!), but, his descriptions of gout (as distinct from rheumatism), hysteria, dropsy, syphilis, and “Sydenham's” chorea were and remain true classics. “He possessed the artistic instinct of the great clinician, the power of seeing what is essential and of passing by immaterial details, and, as we have already seen, with him this was a conscious purpose, a programme, a method” [Faber, 1923, p. 11]. Sydenham's friend John Locke (1623–1704), also trained as a physician, considered by some the founder of empirical philosophy, strongly supported Sydenham's approach.

And it was the merit of another brilliant English physician (pediatrician), Sir Archibald Garrod (1858–1936), to document familial occurrence of specific biochemical disorders in humans such that the pioneering British Mendelist, William Bateson, was able to infer, correctly, autosomal recessive inheritance of alkaptonuria [Bateson and Saunders, 1902, vol. I, footnote p. 133–134] basing himself on familial occurrence in children born mostly to normal parents, with high proportion of first cousin unions. In further work, Garrod also concluded on autosomal recessive inheritance in cystinuria, albinism, and pentosuria, the four Garrodian “inborn errors of metabolism” in his Croonian lectures of 1902 [Bearn, 1993].

The word “syndrome,” first used in English in 1541 (OED) is an exact cognate of the post-classical Greek word, but generally used in a causal, not a merely symptomatic sense (e.g., flu syndrome, nephrotic syndrome, non-immune fetal hydrops, sudden infant death syndrome, etc.)

During early post (1900)—Mendelian years segregating traits were initially referred to as “unit characters” or “unit traits” [Castle, 1905, 1906], inferring that the obvious phenotype (e.g., hooded rat, yellow mouse, Mendel's pea phenotypes) was the only trait segregating, perhaps inspired by Mendel's insistence on “pure lines.” However, soon thereafter, it became obvious that “unit characters” were a delusion; hence, the introduction by Plate [] of the concept of pleiotropy: “Production by one mutant gene of apparently unrelated multiple (or manifold) effects at the phenotypic level (polypheny)” [Rieger et al., 1968]. In Plate's own words“I designate as pleiotropic an entity if several traits depend on it … always occur together, thus appear correlatively united.” A solid demonstration of the validity of the concept is provided by Dobzhansky and Holz [1943] in a paper referring to work by Dobzhansky done almost 20 years earlier on fruit fly mutants differing from each other and from normal in the structure of their genitalia and noting that mutant (fruit-fly) strains “kept in laboratories often differ not only in the mutant genes which produce the obvious external traits for which these mutants are named but also in minor modifying genes which affect various structural and physiological characters of the flies” [Dobzhansky and Holz, 1943]. For measurement of a structural trait, developmentally apparently unrelated, they developed a spermatheca index, varying from 1.24 to 1.77 by dividing maximum diameter by maximum height of the spermatheca, a paired chitinized organ of the reproductive tract at the top of the uterus and below the common oviduct. Dobzhansky and Holz [1943] found that most of the 19 mutants they studied and preserved (alleles of white, yellow, ruby, vermillion, forked, and dusky) were associated with pleiotropic changes of the spermatheca: All 4 at the white locus, 9 of 10 yellow locus mutations and the vermillion mutant all modified shape of spermatheca, thus are pleiotropic genes. These authors comment: “A ‘character’ is obviously an abstraction, though unavoidable for descriptive purposes. Genes produce not characters but physiological states which, through interactions with the physiological states induced by all other genes of the organism and with the environmental influences, cause the development to assume a definitive course and the individual to display certain characters at a given stage of the developmental process. … The fact that some mutant genes alter the pigmentation as well as the spermatheca shape makes these genes distinctively pleiotropic” [p. 301]. Hence, it is probably prudent to assume that all gene actions, regardless of single, externally most conspicuous phenotype, may have multiple, more or less cryptic other, pleiotropic manifestations. Vogel [1961] said: “In human genetics (Erbpathologie) [Pleiotropie] is extraordinarily common [not only] in the … hereditary syndromes. But on closer inspection the genes for most other hereditary diseases also turn out to be pleiotropic” (p. 372, transl. J.M.O.). Does a separation therefore between syndromal and “non-syndromal” forms of X-linked mental retardation lack compelling logic? The more Mendelian research has gone into depth, the more examples have become known which can only be understood under the assumption of pleiotropic factors. We saw above that all Y-races of mice have black eyes, in contrast to the y-kind with red eyes. Thus, this allelic pair affects pigment of fur and of retina. Furthermore, B not only evokes the dark granular hair pigment, but in the presence of CY… also black pigmentation of the skin of the ears, wherefore all wild-colored and black mice alone have gray (rather than white) ears” [rather verbatim translation of Plate's atrocious German of , J.M.O.].

Two comments:

The original noun was Pleiotropie not Pleiotropismus; hence Grüneberg's use of “pleiotropism” is incorrect; 1961 uses the latter term occasionally, and despite acknowledgment of the correct term in print, McKusick continued to use “pleiotropism” consistently.

Secondly, Plate was incorrect in stating “always occur together”; various components of pleiotropic traits can have different penetrances from 0% to 100%, hence are statistically correlated.

Pleiotropy can be conceived of as a consequence of genic parsimony on two levels: (1.) the individual, and (2.) the species.

  1. The individual: “Multitasking” of genes, for example, Fgf8 at work in 72 hr chick embryo in the distal-most limb bud ectoderm, in somitic mesoderm, branchial arches, mesencephalon, neural retina (optic cup), and tail demonstrated by in situ hybridization techniques.
  2. The species: Fgf8 is also used to “fashion” bat wings; it interacts with cranial neural crest to form the mammalian face and telencephalon, normal and regenerating amphibian limbs, and cardiogenic mesoderm. Also, Pax and Hox genes are found in all animal phyla; Pax6 of mouse functions not only in lens and retina but also in neural tube and pancreas and activates the somatostatin gene in the rat.

Hadorn [1955] distinguished between mosaic (genuine), relational, cell-reactive and induced gene pleiotropy.

  • Mosaic pleiotropy due to differential action of the same gene in different cytoplasms resulting in a pattern of mutant gene manifestation which consists of a number of “independent” autophenes, that is, component manifestations of the syndrome, for example, the syndrome of hypothalamic hamartoblastoma, imperforate anus, polydactyly, bifid epiglottis, laryngeal cleft, central polydactyly (Pallister–Hall syndrome) due to autosomal dominant frameshift and nonsense mutations of the GLI3 gene allelic to Greig syndrome [Biesecker, 2008].
  • Relational (spurious) pleiotropy: Mutant gene acts directly on only one cell type but exerts indirect influence on other cells (secondary allophenes) presumably through diffusible substances such as hormones, amino acids, etc.

Mosaic and relational pleiotropy reflect intercellular mechanisms; the following two reflect intracellular mechanisms.

  • Cell-reactive pleiotropy: a mutant gene exerts identical primary actions in different cell types or systems but different characters arise owing to their differential reactions.
  • Induced gene pleiotropy: the primary action of a mutant gene is not identical in different cell systems because the developmental factors are different in these systems giving rise to different characters [Rieger et al., 1968 and 4th ed, 1976].

Other definitions:

King and Stansfield [1997]: “Pleiotropy: the phenomenon in which a single gene is responsible for a number of distinct and seemingly unrelated phenotypic effects.”

Rédei [2003]: “Pleiotropy: one gene affects more than one trait; mutation in various elements of the signal transduction pathways, in general transcription factors or in ion channels may have pleiotropic effects. The existence of pleiotropy was questioned with the emergence of the one gene—one enzyme theory. It was inconceivable on that basis that one tract of DNA could code for more than a single function (“Pleiotropism non est… that is the dogma,” p. 161 in Genetics. 1959. Sutton E.H., ed. Josiah Macey Found, New York). It has been shown; however, that mutation at different sites within single mitochondrial tRNA genes may lead to several different human diseases. The complete sequence of the Drosophila genome shows that ∼13,601 genes encode ∼14,113 transcripts, indicating a minimum of nearly 4% of the genes display pleiotropy. Antagonistic pleiotropy claims that evolution does not work against variations, which adversely affect individuals after the completion of the reproductive stage of life, and the alternative genotypes display opposite phenotypes.”

Initially the term pleiotropy applied only to Mendelian mutations; however, in recent years it has been applied more broadly.

Thus, there is the pleiotropy of:

  • Autosomal dominant mutations—several mechanisms (e.g., Pallister–Hall syndrome).
  • Autosomal recessive mutations (e.g., Meckel, Bardet–Biedl, and Fanconi syndromes).
  • X-linked dominant, semidominant, recessive and male-lethal mutations (e.g., Addison disease, cerebral sclerosis; Albright hereditary osteodystrophy; incontinentia pigmenti; ATRX; Alport, etc.).
  • Mitochondrial mutations: Leber optic atrophy; Leigh syndrome; Kearns-Sayre, Pearson syndrome, etc.
  • Aneuploid pleiotropy (pleiopleudy), now including microdeletions/microduplications.
  • Epigenetic pleiotropy or inheritance—Wiedemann–Beckwith, Prader–Willi/Angelman syndromes; heritable or non-heritable changes in gene expression.
  • Teratogenic pleiotropy—alcohol, jervine, cyclopamine, 13-cis-retinoic acid, etc. may produce holoprosencephaly (HPE) in different species anatomically identical to HPE produced genetically (SHH, DHCR7 mutations), hence, probably interfered with the same signal transduction pathway involved in normal forebrain/facial development.

Stabilizing selection, say of the number of cervical vertebrae (7) in almost all mammals, may involve a strong coupling of such stabilization (against homeotic changes in this number) with deleterious pleiotropic effects [Galis, 1999]. This selection appears to be indirect and is reflected in the huge variety of deleterious morphological and genetic effects associated with a posterior homeotic transformation of C7 into a T1-like, rib-bearing vertebra. We find more or less well developed cervical ribs in almost one-half of all stillborn fetuses [Furtado et al., 2011]. They have also been seen in childhood cancers [Schumacher et al., 1992]; single umbilical artery may be another such sentinel anomaly.


“Induktion ist nichts anderes als Feldwirkung”

Spemann [1936]

“Induction is the same as ‘field action’.”

Idem, 1938, Boedicker et al. translation.

The developmental field as a morphogenetic unit (of the embryo) was a concept conceived independently in 1921 by Hans Spemann and Alexander G. Gurwitsch. The above quote is from Spemann's Silliman Lectures held at Yale University in 1933, published in German in 1936, and translated a.o. by Jane Oppenheimer in . The phenomenon was first demonstrated in the dorsal lip of the amphibian blastopore at Spemann's suggestion by Hilde Pröscholdt (later Mrs. Mangold). Her doctoral thesis [Spemann and Mangold, 1924] based on this work and subsequent publication with Spemann led to a Nobel Prize for Spemann in 1935. His Nobel lecture details this study and cites work from his lab and by others on the several fields discovered to that time [Spemann, ]. Hamburger [1988], also a Spemann student, later at St. Louis University, painted a moving portrait of Hilde Pröscholdt/Mangold at that time and her tragic death at young age shortly after her doctoral promotion and birth of her son who also died young in WWII. Far from slighting Hilde Mangold's contribution to this momentous discovery Spemann kept a large portrait of her in his office as can be seen in the Spemann section of the Museum of Freiburg I (J.M.O.) was once privileged to visit with Dr. Anita Rauch.

I (J.M.O.) once had a similar privilege in Moscow during Soviet days to meet Professor L.V. Beloussov, a grandson of Gurwitsch and himself a distinguished embryologist, to revisit his grandfather's field concept [1922 and 1927] then understood as exerting its morphogenetic effect through extrinsic forces, elegantly modeled in mathematical terms [Beloussov et al., 1997]. In his discovery Spemann was vindicated, but the Gurwitsch formulation was ignored by subsequent workers in embryology as a metaphysical construct. War and subsequent communist repression made it impossible for Gurwitsch to publish amendments and corrections to his theory until it was too late.

Re Gurwitsch Spemann [1936] writes: “If I understand him correctly, the field is largely independent of the material embryo… To fields maybe attributed at times complex anisotropy without deriving it from the corresponding field substrate.” (Transl. J.M.O.).

Paul Weiss, then working in Vienna (later at the University of Chicago) derived his field concept (Das Determinationfeld, 1925) from Gurwitsch but based it on results of amphibian regeneration. See also Weiss 1925, 1926, 1928. The subsequent formulation of his field concept was most clearly and extensively expressed in his textbook [Weiss, 1939] and reviewed by Oppenheimer [1967].

It was once pointed out that nowhere in his magisterial text does Spemann ever mention the gene. And it was the inability of subsequent workers to identify biochemically the “stuff” that makes the field work, that is, exert its inductive effect, and the progressive disciplinary alienation between geneticists and embryologists (who, after all were expected to provide the Rosetta stone on the relationship between evolution and development as due to changes in embryogenesis) that kept the “field” a developmental, not a causal concept, and led to its virtual intellectual oblivion, surviving only here or there as inscrutable incunabula [Gilbert et al., 1996; Opitz, 2012]. We are still alive to testify that, in part, it was clinicians (!) who breathed life again into the developmental field concept and made a truly causal entity of it based on the observations of heterogeneity, homology, and phylogeneity.

Here only with respect to heterogeneity: If the same malformation, say holoprosencephaly, is shown to be due to 2 or more different causes then it evidently constitutes a dysmorphogenetic unit of the embryo reacting, because of developmental constraints, in like manner to, say, Mendelian mutation, aneuploidy, and environmental causes. This anatomical complex then must also constitute a morphogenetic unit under normal conditions responsible for the production of “morphogens” qua “form producers” in Turing's [1952] terminology. Already in the first edition of his text [1985] Scott F. Gilbert, reviewing the work of Whittaker on tunicate development, concluded “that certain determinants that exist in the cytoplasm cause the formation of certain tissues.” These… morphogens “appear to work by selectively activating (or inactivating) specific genes. Thus, the determination of the blastomeres and the activation of certain genes are controlled by the spatial localization of morphogenic determinants within the egg cytoplasm.”

Carlson [2003] discusses the morphogenetic field as follows: “… a region of the body in which the cells contain the information sufficient to form a specific structure, such as a limb or an eye. Such a region has boundaries which often shrink in the course of development; and in contemporary parlance, a field may be a region in which a selector gene or specific combination of homeobox genes is expressed. Within an active morphogenetic field, the cells in aggregate possess certain properties, such as regulation. On an individual basis, there appears to be considerable room for autonomous behavior among cells within a morphogenetic field, and this property has been viewed as one in which individual variation can be expressed.”

The fact that there exist many individual malformations of different parts of the body without substantial impairment of the morphologic integrity of the rest of it strongly suggests that developmental fields have a certain limited degree of developmental latitude, thus, perhaps also autonomy to change ontogenetically in response to evolutionary pressures. How else to explain in more or less closely related species: long, short, no tail; huge or humble canine teeth; fully functional versus absent 12, 13, or 14 pairs of ribs; male genitalia with or without baculum or penoscrotal inversion; web or no web between toes; hooves, nails or claws (retractable, non-retractable)…etc. This is perhaps what Meckel [1812] had in mind when he wrote: “Die ursprünglichen Fehlbildungen sind nicht wider [sic] die Natur.” (Primary malformations are not contrary to nature.)

One may therefore wonder whether human organs that are malformed in Meckel's sense of normal states in other species, should not be regarded as actively undergoing evolutionary modification.

Hence, also the concept of modularity [Gass and Bolker, 2003] viewing organisms as the “integration of partially independent, interacting units at several hierarchical levels.” “The key feature [sic] of an evolutionary module is its ability to change independently of the rest of the body.” [Gass and Bolker, 2003]. Fields therefore are also modules; the adult body of metazoans is a mosaic of modules arising from a mosaic of developmental fields in the embryo. This is in contradiction to Cuvier's “correlation of parts” making the body an indecomposable, tightly integrated whole so that tinkering with one part or “cog” in it has the entire organism breaking down like a house of cards or a clock with broken mainspring. Schlosser and Wagner [2004] review summarizes nicely “the role of modules in development and evolution.” Contrast the above definition of fields with Raff's [1996] definition of modules: “They have an autonomous, genetically discrete organization… they contain hierarchical units and may in turn be parts of large hierarchical entities… they have physical locations within the developing system… [with] varying degrees of connectivity to other modules… [and] developmental modules can undergo temporal transformations.”

Also of note is the authoritative discussion of modularity in Gilbert [2010]. The re-assessment of human (primary) malformations in comparative modular terms may contribute enlightenment as to our recent evolutionary history.

To return to embryonic fields, morphogenetic processes within fields may be considered:

  • Epimorphically hierarchical (from primary field to progenitor field to final, secondary, or epimorphic fields).
  • Epigenetically complex, nonlinear, and reciprocally interactive but not so interactive as to bring down the entire organism, as Cuvier worried, when evolution begins to tinker with one part;
  • Spatially delimited and coordinated;
  • Temporally synchronized, and
  • Phylogenetically highly constrained into a few final, common developmental paths, normal (many genes, few structures—compared to dogs or fancy pigeons, humans are an anatomically dull lot) and abnormal (even more mutations, yet HPE (or cleft palate, etc.) occurring over and over again, not only in humans but in so many other vertebrates).

And if a specific teratogen (e.g., alcohol or jervine/cyclopamine derived from Veratrum californicum) leads to a corresponding malformation, such as HPE in different species morphogenetically identical to one of genetic origin (SHH mutation), then it must have interfered with the same basic signal transduction pathway, in this case cholesterol modification of the sonic hedgehog protein [Opitz, 2012].

The first point above requires some elaboration. Before the molecular era we viewed development as progressing from primary (organizer) field to the final, secondary, epimorphic fields with no transitional stage. Thus, it was the great merit of Eric Davidson to have conceived of the progenitor field. His formulation [Davidson, 2006, pp 119–120] is as follows:

Progenitor field specification in Bilateria occurs in canonical locations in respect to the body plan: brains form at the anterior end, hearts in the paraxial mesoderm, the intestine is posterior to the pharynx, in vertebrates the appendages occur at branchial and pelvic levels of the trunk and so forth. These locations are defined with respect to embryonic spatial coordinates. As an output of the late embryonic regulatory system, genes encoding signal ligands are expressed in given domains of the embryo, and the cis-regulatory modules controlling expression of progenitor field regulators respond directly to these signal inputs. For example, the expression of the tinman gene in Drosophila defines the progenitor field for the future heart with respect to the design of the embryo as expression of its ortholog, nkx2.5, defines the heart progenitor field in vertebrates. The tinman cis-regulatory module controlling this phase of expression is motivated by transcription factors (Mad, Medea) which are activated in response to Dpp signaling, and this is the signal, transcriptionally expressed early on in the dorsal portion of the embryonic blastoderm, which designates the dorsally migrating bilateral strips of mesoderm where the heart will form. An additional embryonic spatial input is expression of the Twist transcription factor, which is expressed from the beginning in embryonic mesoderm: so these embryonic inputs in combination specify that the heart progenitor field will start out in the dorsal mesoderm of the embryo body plan.

In progenitor field specification the cells of the progenitor field must be in (relatively) unspecified state before expression of the relevant regulatory gene, and the location of the progenitor field should be malleable, depending on where the signal is expressed. Both have often been shown to be true. This is the general import of observations recurrent in the long history of vertebrate experimental embryology, in which transplantation of tissue (which proves to be the source of the signals requisite for progenitor field specification) was found to cause ectopic body part development. Modern studies in the chick, for example, have established that ectopic presentation of Fgf signal ligands can induce ectopic limb formation anywhere in the lateral region between the normal limbs, the forelimb and hindlimb buds are distinguished by different transcriptional regulatory states (i.e., expression of ptx1 and tbx5, respectively), though earlier both express hoxc9 and other posterior group hox genes. Primary sources are Davidson [1993] and Davidson and Erwin [2006].

Even a casual glance into the magisterial volumes edited by 2001, now Hennekam, and by the late Charles Epstein and his collaborators on Head and Neck Anomalies, and “Inborn Errors of Development,” [Epstein et al., 2008] respectively, will suggest that malformations are defects of signal transduction cascades, or pathways in developmental fields and their overall consequences for structure and function (developmental programs). They may be due to mutations in the coding sequences (DHCR7 mutations in RSH/SLO syndrome) or regulatory defects (e.g., chromosome translocations on 17q upstream of SOX9 without disruption of the gene but still causing a form of campomelic syndrome). They may involve the primary field (heterotaxies due to mutations of ZIC3, ACVR2B, EGF-CFC, LEFTAA, EBAF, etc.); progenitor fields (holoprosencephaly due to mutations of SIX3, SHH, TGIF, ZIG2, PTCH1, GLI2, D1SP1, NODAL, FOXH1, etc. [see Roessler and Muenke, 2010], or secondary epimorphic fields (syndactyly of toes 2/3, rudimentary postaxial polydactyly, submucous cleft of palate, mild hypospadias, etc.).


Since its inception, the concept of “Associations” in clinical genetics has engendered feelings of unease and vagueness; unease because statistics cannot make biology only help biology in assessing likelihood of chance concurrence of various developmental effects; and vagueness, because the concept lacks causal underpinnings. Some (pediatric) house officers ask: “What is the difference between an Association and an association?1 Why is, for example, the VATER association not a syndrome?”

It was Isidore Geoffroy St-Hilaire (1805–1861), the son of Étienne (1772–1844, and 40-year friend of Cuvier, [Le Guyader, 2004] who was apparently the first to wrestle with the concept of “Associations” in his monumental 4-volume Traité de Tératologie [1837], distinguishing between coincidental concurrence: simple recontre or simple juxtaposition (Günther's Zufallssyndromie) and “combination of anomalies”: intimement associée, not une seule et unique anomalie but une anomalie complexe in the sense of a polytopic field defect, for example the DiGeorge complex.

In the 20th century (von) Pflaundler and von Seht [1922] attacked the problem statistically without reference to I. Geoffroy St-Hilaire's [1837] work on anomalies associées. Since Pfaundler was chairman of Pediatrics and of the Children's Hospital of the University of Munich, he had command of patient material in the thousands. His syntropy index (s), referring to two conditions A and B was: s = ((Pab)/(Pa·Pb)). Thus, Pfaundler and von Seht calculated s for acute joint inflammation/cardiac involvement an index of 58 (statisticians would call that result “highly significant”), for joint inflammation/chorea = 12.9, for chorea/rickets = 0.09 (“negative” syntropy or association). Günther [1948] and Kolb [1957] produced later refinements (available in translation from J.M.O.).

Quan and Smith [1973] following upon the earlier observation of Say and Gerald [1968], studied corresponding cases at The Children's Orthopedic and University Hospitals in Seattle, WA between 1964 and 1970. Their Case 1, a girl, had tracheo-esophageal fistula (TEF) with esophageal atresia, anal atresia with rectovaginal fistula, short right first metacarpal, ectopic right kidney, absent pedicles of the 5th lumbar vertebra, absent left 12th rib and 10th thoracic hemivertebrae. Case 2, also a girl, had TEF with esophageal atresia, anal atresia with rectovaginal fistula, “bifid” 3rd and 4th thoracic vertebrae and right radial aplasia with absent thumb and left hypoplastic thumb.

Quan and Smith [1973] comment: “The probability of the simultaneous occurrence of any of these defects in the same individual, based on their individual incidences, is so unlikely that it suggests a non-random association.”

The references on the basis of which frequencies were calculated were: Kaufman et al. [1970], Tünte [1968], Pardini [1968], Moore and Lawrence [1952], Conway and Bowe [1956], Ladd and Gross [1934]. See also Källén et al., 2001.

Methodologically, Botto et al. [1997] on VATER have the last word, concluding also that their data “offer evidence for the specificity of the VATER association… and raise the question of a common pathway for patterns of VATER and other types of defects in at least a subset of infants with multiple congenital anomalies.” Thus, Botto et al. [1997] have brought the concept of association at least to a pathogenetic, if not a causal, level.

The International Working Group [IWG, Spranger et al., 1982] meeting in Berlin at the VII International Congress Human Genetics in 1986 stated that “associations are real biological entities of great importance to the clinician because of their pathogenetic and genetic counseling implications.” But concluded that a syndromal approach to associations was unproductive and perhaps even incorrect since associations are causally non-specific entities [ital. added, Evans et al., 1987].

It is remarkable that there has been virtually no increase in understanding of associations since then (1986) except to reiterate that they are:

  • Real biological entities.
  • Do not segregate as Mendelian entities.
  • Chromosomally apparently normal.
  • Sporadic with negligible recurrence risk, except for prior spontaneous losses [Martínez-Frías and Frías, 1997].
  • Highly lethal, hence are seen not uncommonly in fetal and pediatric pathology services.
  • Appear to be multiple defects of blastogenesis; hence, one might be tempted to view them as apparently idiopathic multiple anomalies of blastogenesis (i.e., gestational age days 0–28).

Example: our (Utah) autopsy PCA 000 of 2012: Male fetus born at 27 weeks gestational age (GA) to a 19-year-old primigravid woman, weighing 734.1g; died neonatally. Prenatal ultrasonography at 25 weeks GA showed multiple anomalies and anhydramnios. Autopsy showed (Fig. 4): Sirenomelia, bilateral renal agenesis with Potter sequence, absent external genitalia and anus, atresia of sigmoid colon, microtia (left), tetralogy of Fallot, right aortic arch with retro-esophageal left subclavian artery, retro-aortic innominate vein, absent ductus arteriosus, unilobate left lung, duodenal atresia, single umbilical artery; gross absence of thymus, but microscopic evidence of cervical thymus and parathyroid with apparently normal chromosomes, no evidence of 22q11.2 deletion or duplication. Evans et al. [1987]: “The use or [sic, “of”] the term ‘associations’ seemed to grow out of attempts to represent teratological vulnerabilities that could be seen under a variety of circumstances… [and] that associations actually represent embryological relationships that determine which developmental fields are likely to be affected by a particular teratogenic agent or event.” Since associations seem to be coded on birth certificates and stillbirth certificates in Hungary, Czeizel in Evans et al. [1987] was able to estimate their prevalence in the Hungarian Congenital Malformations Registry between 1973 and 1982. Total prevalence was 1.22 association cases per 1,000 total live and stillbirths (1 in 820); associations representing almost 29% of total multiple congenital anomalies (MCA) cases and 41% of those with unidentified MCA patterns. Evans et al. [1987] conclude that: “Associations may very well be useful indicators of general reproductive well-being in populations.”

Figure 4.

Fetus with presumed association at GA 27 weeks, weighing 734.1g. A: Potter sequence (lanugo of forehead normal for population group.) Right lateral view—normal ear (flattened). Left lateral view: Microtia. B: Sirenomelia. C: Left hand: Small thumb, hypoplastic thenar eminence (mild radius dysgenesis sequence) and abnormal palmarflexion creases. D: Abnormal heart: Tetralogy of Fallot, right aortic arch, retro-aortic innominate vein, absent ductus arteriosus. E: Marginal insertion of umbilical cord with single umbilical artery; single lobe of left lung. 22q11.2 region normal.

Further studies to elucidate causes of associations might be:

  1. Using the resources of the Utah Population Database attempting to determine if any or all “association” cases in Utah may be related, more or less closely or distantly, to each other, that is, can they be traced to one or a few ancestors. If so, that may indicate genetic predisposition, dominantly inherited but with threshold which, in a given case, may be lowered below the population level by adverse circumstances. A model of this is Sewall Wright's otocephaly work [1928] or his studies of “atavistic” guinea pig “polydactyly” [Wright, 1934, 1935].
  2. Perform a retrospective study of health and reproductive status of the women in Utah who have had a fetus or infant with an idiopathic association; it would be of great interest if 5–10 years later they had all developed diabetes, or lupus, or cancer, etc. Or, if in subsequent pregnancies they had had an adverse developmental outcome.
  3. Preferably do both studies together.
  4. Perform exome sequencing in those with apparently normal chromosomes.
  5. Lastly, since it has been possible to produce all VATER anomalies in the mouse by disrupting the sonic hedgehog/Gli transduction pathway [Motoyama et al., 1998; Kim et al., 2001a, 2001b] it is suggested to search this pathway most carefully at least in the VATER/VACTERL subset of association infants.
  6. Perform exome/genome sequencing on affected and unaffected organs in subjects with the VATER association. This approach has led to the identification of activating mutations of AKT1 in Proteus syndrome [Biesecker et al., 2011], and also, remarkably, to the appreciation of the role of tissue-specific (mosaic) mutations: mutations were present in Proteus syndrome in affected but not in unaffected tissues (e.g., results of exome sequencing of blood were normal). VATER associations and Proteus syndrome are of course vastly different conditions. Nevertheless, they seem to share some general findings, including sporadic occurrence, low recurrence risk, normal genetic studies in blood, and possibly higher rate of twinning (established in VATER association, unclear in Proteus syndrome). Presumably, based on the greater frequency and malformation patterns in associations, if tissue-specific mutations are found in the VATER association they will likely be inactivating, rather than activating as found in Proteus syndrome, and possibly involving one of several genes, rather than a single gene [Botto, 2013, personal communication].

Minor Anomalies and Mild Malformations

Another task David Smith assigned one of the authors (J.M.O.) upon arrival in Madison in 1961 was to examine all of the newborn infants at St. Mary's Hospital for the presence of minor anomalies, malformations, deformities and the rare syndrome that came along, for example Down syndrome, Klinefelter syndrome—(micro-orchidism) and the eventual Zellweger syndrome. The study of Marden et al. [1964] stimulated a productive world-wide confrontation with the prevalence, definition and pathogenesis of minor anomalies and their predictive value for the presence of undetected internal anomalies, potential syndromes and childhood malignancies [Opitz, 1985]. Studies a.o. in Sweden [Ekelund et al., 1970], Hungary [Méhes, 1983, 1988] the USA [Hook et al., 1976; Leppig et al., 1987], Israel [Merlob et al., 1985] signified not only confirmation of Marden et al. [1964] in general terms but also a dramatic shift of such concerns and methodologic approaches from anthropology into medical genetics. Causal concerns were always on top of the agenda, that is: to what degree might single transverse creases, fifth finger clinodactyly, distal palmar axial triradii, radial loops on digits 4 and 5 (all minor anomalies, per se of trivial concern to the infant, parents totally oblivious to them) be a miner's canary for something far more serious, for example, Down syndrome?

The Israeli approach ultimately was the most fruitful, being extended to an analysis of fluctuating asymmetry (FA) as an exquisitely sensitive indicator of developmental stability/instability of anthropomorphic traits not grossly evident to the clinicians as a “minor anomaly,” say in a premature infant [Livshits and Kobyliansky, 1991; Livshits et al., 1998]. Since apparent minor anomalies may be nothing more than a normal variant in a family but unusual for the population, the phenotype analysis of such findings in an infant at birth requires first-class epistemological judgment, preferably an analysis of family resemblance if the matter warrants concern. I for one avoid “dysmorphic features” as much as “etiology” – the former does not discriminate between minor anomalies and mild malformations, and the latter not between cause and pathogenesis.

Mild malformations may include 50% syndactyly of toes 2 and 3, postaxial 6th finger tag, cleft uvula or cleft xiphisternum, Meckel diverticulum, persistent left superior vena cava, grade 1 or 2 hypospadias, etc. Conventionally they are viewed as defects of organogenesis, the second half of morphogenesis, days 29–56, from the end of blastogenesis until the time of metamorphosis from embryonic to fetal life. They are more likely monotopic than the polytopic defects of blastogenesis, less often lethal or associated with (MZ) twinning, more often show skewed sex ratios (apparent indicator of multifactorial determination), less strikingly confined to midline, and are more likely autosomal dominant traits. Some apparent mild malformations (say in an infant of a diabetic mother) may, in fact, be defects of blastogenesis recanalized or re-equilibrated toward more normal development [Waddington et al., 1947]; in the early part of the 20th century many experimental embryologists showed that the embryo has a remarkable capacity of repair and re-regulation of development, even after drastic manipulations. Defects of blastogenesis may occasionally be associated with tumor (teratoma) formation; there is increasing evidence that defects of organogenesis may be associated with later cancer formation, best documented in the cancer risk “associated with” cervical ribs [Schumacher et al., 1992; Merks et al., 2005].

Phenogenesis [Fischer, 1939 fetal development, gest. age day 57–266]: Growth, maturation, histologic differentiation, and attainment of all of those quantitative and qualitative characteristics that identify the fetus as member of a species, population group, and a specific family. Failure by family to identify the latter (no family resemblance) is a strong indicator of aneuploidy. Dominant inheritance of a group of minor anomalies may indicate segregation of a microduplication (less likely—deletion).

Abnormalities of phenogenesis are minor anomalies to be distinguished, in most cases, from mild malformations, that is, defects of blastogenesis and of organogenesis.

Terminology of minor anomalies and of mild malformations was recently reconsidered, to some extent reformulated, and promulgated by an international working group [Allanson et al., 2009]. To generalize, cautiously:

  • Mild malformations are all-or-none, qualitative, binary traits; minor (“anthropometric”) anomalies are quantitative traits with specific Gaussian frequency distributions, depending on age, sex, and population group.
  • Mild malformations may be dominantly inherited; are common traits in the population with little/no selective disadvantage, that is, no impairment of fitness. Minor anomalies are, by definition, multifactorial/polygenic traits and “obey” Galtonian inheritance (q.v. any anthropology text).
  • Mild malformations (e.g., absence of palmaris longus muscle, cleft uvula…) generally do not alter family resemblance; (multiple) minor anomalies may alter or abolish family resemblance.
  • Multiple mild malformations may represent Mendelian pleiotropy… multiple minor anomalies: aneuploidy (beware of over-interpreting population variants as minor anomalies; always try to do the phenotype analysis as an analysis of family resemblance, that is of sibs and parents at a minimum).
  • Mild malformations are assessed objectively on “surface” examination; but an apparently normally developed fetus at autopsy may have an astonishing inventory of mild malformations inside (laryngeal cleft grade 1–2, high bifurcation of carina, defect of pulmonary lobation, non-fixation of mesentery, malrotation of gut, horseshoe kidney, uterus duplex, single umbical artery…). Minor anomalies are assessed anthropometrically—broad/long head, hypertelorism, short fingers…. but frequently are subjective traits (as a mother of a girl with C-trigonocephaly syndrome from Buenos Aires said: Her second affected stillborn child, a boy “was as ugly” as the first-born affected child, a girl).
  • We have the impression that mild malformations do not show ethnic differences in prevalence, but lack concrete data, also on whether minor anomalies vary in prevalence by “race.”
  • As pointed out many years ago by Marden et al. [1964], multiple minor anomalies may be an indicator of co-existing malformation (think of all of the minor anomalies in Down syndrome as potential indicator of AV canal, in Ullrich–Turner syndrome to predict pre-ductal coarctation and/or horseshoe kidney…).
  • Malformations may cause minor anomalies (hypotelorism in HPE, dermatoglyphic defects in radius dysgenesis or ectrodactyly).
  • All apparent minor anomalies may occur as normal variants in the population; multiple normal variants may at first glance impress as multiple minor anomalies—an entity that may be called a “VFDP”—Variant familial developmental pattern, to be determined as such only after an analysis of family resemblance!

A study of dermatoglyphics of fetuses, families, and twins in different population groups and different conditions with correlation coefficients, analysis of variance and of anatomy is highly instructive with respect to normal and abnormal human variation; think only of the pioneering studies of Midlo and Cummins, Bonnevie, Penrose, Sarah Holt, Irene Uchida and so many others defining Down syndrome (low total ridge count, distal axial triradii, preponderance of ulnar loops and of radical loops on 4 and 5, third interdigital distal loops, hallucal open field); 18 trisomy syndrome (10 low arches); Ullrich–Turner syndrome (as many as 10 whorls with high total ridge count of 160 to over 200), etc.

But, reality is messy. Consider the lad from Israel with evident FG syndrome [FGS1 molecularly confirmed, Graham et al., 2010] with a right broad thumb (a minor anomaly) and a left bifid thumb (a malformation) requiring amputation. We all have had similar experiences in individuals with small and contralateral absent thumb or, small/triphalangeal thumb, small auricle/microtia–anotia, small/absent upper lateral incisors/broad/bifid upper lateral incisors, brother with broad/sister with cleft xiphisternum, etc. Thus, it seems appropriate to propose a hypothesis: Bilaterally symmetrical structures of the human body (which are conjoint MZ twins) are subject to fluctuating asymmetry [FA, Ludwig, 1932, 1936; van Valen, 1962; Livshits et al., 1998; Opitz and Utkus, 2001]: random differences between the two sides of quantitative traits. These increase in parallel to the decreased buffering ability of an organism, thus, are a measure of developmental (in)stability: The greater the FA the less the developmental stability. Developmental buffering refers to the “amazing stability of complex developmental processes” to achieve a standard phenotype, one with low variance about the mean, in spite of genetic and/or environmental disturbances [Schmalgausen, 1949; Waddington, 1940, 1957; Gibson and Wagner, 2000]. This is the paradox of evolution: Transmission of phenotypes over many generations with great fidelity and little or no impairment of fitness, and yet with sufficient recombinant/mutational variability between generations to allow for the action of natural selection. Artificial selection, for example for large body size in dogs, increases FA and probability of polydactyly [St. Bernards, Great Pyrenees vs. Poodles, Alberch, 1985]. Primates offer a particularly striking insight into the evolution of digit structure [q.v. Golden Potto—Arctocebus calabarensis, Lemur—Lemur macaco; Spider monkeys, Ateles geoffroyi and Brachyteles arachnoides with thumbs absent or “poorly developed,” Hershkovitz, 1977].

It seems therefore that evolution has tinkered extensively with the medial and lateral digit fields (modules) of many species of mammals, genetically best analyzed by Wright in guinea pigs in 1934 [q.v. Wright, 1984]. Wright was able, by inbreeding over many generations, to restore, “atavistically,” a more or less perfect pentadactyl foot/hand of the Eocene (40 MYA) guinea pig (Cavia porcellus) ancestor, suggesting that this reversion occurred on a polygenic basis with thresholds: by assuming semidominance with two thresholds (one for any development, and one for perfect 5th toes), with standard deviation of 80% of the interval between thresholds, and the means of the crosses taken such that the distributions did not cross the lower threshold or fall significantly below the upper threshold. Wright's elucidation of this continuously distributed trait with threshold became the “grandfather” of all subsequent polygenic/multifactorial analyses with threshold [Grüneberg's “quasi-continuous variations,” Grüneberg, 1952]. Remembering the large environmental component (81%) in this multifactorial system it is not surprising that genetically identical littermates either had normal guinea pig toes (4,4/3,3, lacking thumbs, halluces and 5th toes) or were “polydactylous.”

Thus, it would seem quite reasonable to postulate broad thumbs in FGS1 as a polygenically distributed trait with threshold such that one thumb may merely be broad, the other bifid. “The rate of evolution under natural selection is proportional to the amount of the additive variance” [Fisher, 1930].

It would appear also that in initial stages of morphogenesis the embryo deploys primarily batteries of oligogenes as in numerous parallel, partially overlapping signal transduction cascades; but, that morphological fine-tuning and finishing of the fetus during metamorphosis and phenogenesis require ever more complex polygenic/epigenetic buffering systems. Functioning, under normal circumstances, reliably enough to ensure, more or less, morphological and functional fidelity sustaining survival of the species, but allowing enough wobble in the semi-autonomous systems (modules), will allow action of natural selection and eventual emergence of new species. Also a reason why in small embryos, say of Down syndrome, no minor anomalies, only gross malformations are evident, with the former appearing only gradually as the fetus matures toward birth.


To Lindy Stokes for expert document preparation, logistic support, reference research and illustrations. And to Dr. Jürgen W. Spranger and Dr. Lorenzo D. Botto for their gracious and most helpful discussions of this work, and to colleagues worldwide who have commented on our efforts in biology and the history of medical genetics over the years.


  1. 1

    And some of the German pediatricians of Milwaukee used to joke: “And what, pray tell, is the difference between a VATER and a MUTTER association?”