Background

Mammals are a diverse group in which to examine development and evolution, and besides the mouse and the rat used in biomedical research, provide subjects based on which experimental (Harjunmaa et al. 2014; Montandon et al. 2014; Parsons et al. 2015) and comparative (Cooper et al. 2014) studies have provided major insights. They can effectively serve to illustrate evolution in the classroom and to the public at large given the interest and familiarity of people with the group (Prothero 2007; Asher 2012).

The mammalian morphological diversity, also called phenotypic disparity, is large and encompasses forms as different as bats, whales, mice, humans, the egg-laying echidna and platypus, and the kangaroo. This disparity emerges through the evolution of organogenesis, the portion of individual development in which the general ‘body plan’ as well as species-specific features emerge (Gilbert 2013), followed by the growth process. In placental mammals, organogenesis takes place mostly in the uterus, whereas in monotremes and marsupials a very immature hatchling or newborn, respectively, develops further either close to the mother or in its pouch (Werneburg and Spiekman in press). Many features can be used to characterize developing mammals externally, and each can evolve. Among them are aspects of the integument such as hair, the limbs, and structures of the head such as the external ear or the eyelids (Schoenwolf 2008; Werneburg and Sánchez-Villagra 2011; Werneburg et al. 2016).

The exercise presented here deals with an aspect of development that although not trivial, does not require a rich anatomical background, as a brief introduction using pictures and drawings of embryonic series can easily allow students to extract basic information on external organs. The use of pictorial documentation is tied to a fundamental aspect of anatomical research, one with deep historical roots. During the golden age of comparative embryology, around 1900, hundreds of embryos were illustrated in beautiful treatises that showed different stages in the ontogeny of a species (Hopwood 2005, 2007). The plates are transformation series occurring in the life of an animal. Different organs first appear while others differentiate, so that the species-specific features arise gradually, in a particular sequence of events.

The exercise consists of the comparison of developmental series of diverse species, as documented in illustrations from which anatomical information can be extracted. It deals with features of embryology of vertebrates treated in some works in developmental genetics, but from a whole-organism perspective. The latter concerns mostly the ‘pattern’, referred to as the phenomenological aspect of development (Hanken 2015), as opposed to the approaches that aim at discovering the ‘processes’ or mechanisms behind those patterns (Cubo 2000; Richardson et al. 2009). The organismal perspective (Maier 1999) has a long tradition and involves sophisticated methods of morphometrics and knowledge of anatomy. Current curricula at university and high-school level tend to concentrate on molecular tools and genetics to the detriment of organismal biology.

The comparison of developmental series serves as an introduction on how differences among adults—how morphological transformations in evolution—are the result of developmental repatterning. Repatterning can concern timing (heterochrony), space (heterotopy), quantity (heterometry), and kind (heterotypy) (Arthur 2011). Examples of these can be seen in an examination of organogenesis. The evolutionary changes in developmental timing, heterochrony, have been a focus of research for decades (Raff 1996; Smith 2001; Maxwell and Harrison 2009).

An exercise on comparative organogenesis in vertebrates

The activity is designed for use in lecture-type courses but is scalable to large courses and can be performed including discussions in 45 min. It is implementable without assistance in a class with as many students as the number of developmental series depictions provided (Table 1; Additional files 1–18) or in multiples of that number should pairs or groups of 3 or more students deal with each species. There are five steps to be followed:

Table 1 Selected normal plates of vertebrate development
Full size table
  1. 1.

    Students are given each a set of drawings of one species, which they are expected to examine in a temporal sequence from early to late, based on the obvious progressing nature of development. In the Additional files 1–18 to this article, we provide plates of ‘normal tables’ of development (Keibel 1897) which can be used for this exercise. All specimens on the plates have numbers that serve to order them. In the case of small student groups and where the logistics permit, it is recommended that after being given the complete plates, students cut out each individual without their number and the sets of the series are exchanged among student groups. That way students then have to establish the order of a new set of individual embryos, having been previously exposed to one complete, ordered series of another species.

    Instead of presenting sets of developmental series uniform in terms of the number of specimens and the time window portrayed, we suggest to provide the students with the plates from the original references (Table 1; Additional files 1–18), as this has the advantage of facing the student with a situation more similar to that encountered in actual comparative embryological work. This means having a different number of specimens for each species. An equal number would provide the false impression that each ‘stage’ depicted is comparable to the corresponding one in the series of the other species. The original plates serve also to illustrate the difficulties of establishing stages, and how each of the original authors (Table 1) had a different opinion on how many specimens best characterize a species’ development, and the different and subjective criteria to identify ‘stages’. This exercise is good training against typological thinking, which has had a negative influence on studies of development and evolution. Evolution is about variation and not about fixed types or archetypes (Richardson et al. 1999; Werneburg 2009).

    As a general reference for our own species, Fig. 1 illustrates a subset of human embryos encompassing approximately the first 2 months after conception. It is recommended that all students examine the human series.

    Fig. 1
    figure 1

    Selected human embryos from the “Normentafel” of Keibel and Elze (1908). Approximate age of embryos based on O’Rahilly and Müller (1987): a 22 days, b 24 days, c 25 days, d 26 days, e 28 days, f 32 days, g 33 days, h–l 37–41 days, m–p around 44 days, q 47 days, r 50 days, s 52 days, t 56 days. Embryos not to scale

    Full size image
  2. 2.

    Students examine and subsequently label several anatomical features and their occurrence in the different stages, either digitally or in provided printouts (Additional files 1–18). For this purpose, the students are provided a reference figure which identifies basic structures such as the eye, somites, tail, lower jaw, forelimb, hind limb, and branchial arches (Fig. 2). Further character descriptions and illustrations can be found at https://en.wikipedia.org/wiki/Standard_Event_System. (access: 2016-11-25). This is basically an exercise in identifying structures that are homologous, and as such emphasizes homology thinking, a central aspect in evolutionary biology that benefits from the tree-thinking perspective (Ereshefsky 2012; Wagner 2016), also central here.

    Fig. 2
    figure 2

    Selected vertebrate embryos of different developmental periods taken (except for h) from the “Normentafeln zur Entwicklungsgeschichte der Wirbeltiere” edited by Franz Keibel from 1897 to 1938 (Hopwood 2007) and selected discrete embryological characters as defined in the Standard Event System (Werneburg 2009; https://en.wikipedia.org/wiki/Standard_Event_System). a Common mudpuppy Necturus maculosus (Eycleshymer and Wilson 1910); b, i roe deer Capreolus capreolus (Sakurai 1906); c Triturus vulgaris (Glaesner 1925); d, j Lacerta agilis (Peter 1904); e Sundra slow loris Nycticebus coucang (Hubrecht and Keibel 1907); f South American lungfish Lepidosiren paradoxa (Kerr 1909); g Spiny dogfish Squalus acanthias (Scammon 1911); h goat Capra hircus (Tsukaguchi 1912); k rabbit Oryctolagus cuniculus (Minot and Taylor 1905). Embryos not to scale

    Full size image
  3. 3.

    The resulting series from step 1, revised after closer examination resulting from step 2, are then placed together and compared. This comparison reveals the commonality in the general pattern of differentiation, but also the differences among species in the sequence of appearance of structures. Likewise, it makes clear how some structures form in some groups and not in others; e.g., limbs in land vertebrates (tetrapods) and fins in fishes or scales in sauropsids and hairs in mammals.

  4. 4.

    After a phylogeny-free first comparison, subsequently the species examined are organized in a provided phylogenetic framework (Fig. 3). The subsequent group discussion is about how patterns emerge (which characters are common and different among species) that can be best explained as determined by evolutionary history (see below).

    Fig. 3
    figure 3

    Phylogenetic framework of the species for which developmental series are provided in the Additional file 1. Many natural history museums still depict evolutionary patterns as ‘orthogenetic’, and thus as a linear and directed sequence from ancestor to descendant, including even the classic example of horses [discussed by MacFadden et al. (2012)]. This kind of representation is wrong, as the pattern is actually a branching one. This mistake communicates antiquated knowledge and perpetuates misconceptions about evolution. People tend to see evolution as a story with a beginning, middle, and an end [discussed by Baum and Smith (2012)]. Phylogenetic trees challenge this view, showing a branching and fractal pattern instead of linearity, and with one beginning and many ends

    Full size image
  5. 5.

    The activity is rounded up by showing the video discussed in the following section of this article.

The exercise is useful in showing depictions of real organisms, the common pattern of development of humans, and their evolutionary relatives. An example of general similarity between evolutionary and developmental transformation (Macrini 2002; Martin and Ruf 2009; Asher 2012; Ramírez-Chaves et al. 2016; Werneburg and Spiekman in press), is the fact that the hand in some stages of mammalian foetuses, including human ones, looks like a paddle (e.g., Fig. 1k) and thus resembles superficially that of our aquatic ancestors. This commonality among species in the transformation series contrasts with the differences in the static stage represented by the adult. Here it is important for the instructor to emphasize to the students that there are no steps in ontogeny, but instead that each depicted embryo represents a single, living individual with features of its own that allow it to survive. The individuals represented in the series are examples of populations, so that not only interspecific but also intraspecific variability occurs (de Jong et al. 2009). Furthermore, a clear definition of characters is always important to make reliable comparisons among the specimens of one developmental series as well as among different species. The detected differences among species highlight the importance of studying embryonic features to understand evolutionary changes. Among other subjects of potential discussion are the relation between developmental and evolutionary transformation, recapitulation, and developmental repatterning.

The idea that ‘ontogeny recapitulates phylogeny’ is widespread among students and the public and it is the classic subject of recapitulation. For recapitulation to happen, an addition at the end of the original or ancestral developmental sequence or trajectory would have to occur (Fig. 4; Wägele 2005). That ontogeny does not simply recapitulate phylogeny is very well accepted. Only specific characters or character complexes, such as gill slits in mammals (Fig. 1e), can be recapitulated and, in that case, always perform a necessary functional task during ontogeny (Werneburg et al. 2013b). However, none of the embryos resembles an adult of any other species, so examination of the provided depictions of developmental series makes the case clearly. Deviations from the hypothetical recapitulatory pattern occur, as Haeckel (1866: p. 300) himself recognized. Features in a developmental sequence of new emerging events can move around or develop at different speeds, one or more of them can be omitted, or a whole new feature can appear. There are different kinds of developmental repatterning (Arthur 2011), and the organogenesis in the exercise and video presented here serve as general introduction to them. The comparisons among embryos of different species serve to explain patterns that suggest heterochrony, heterotopy, and heterometry, some of which foreshadow differences in adult body form (Richardson et al. 1997).

Fig. 4
figure 4

A consistent pattern of terminal addition in the evolution of ontogenetic sequences, as illustrated here for four species in abstract terms, leads to ontogenetic recapitulation. Each species is characterized by a common developmental trajectory consisting of the first step, M1→M1. A new feature is added at the end of the sequence. In this ideal case, species “D”, the one with the most specialized condition, contains in its ontogeny the sequence of evolutionary transformations. But there can be deviations from the recapitulatory pattern. Features in the sequence can move around, one or more of them can be deleted, or a whole new feature can appear. When those changes are of great evolutionary significance, they are thought of as an evolutionary innovation, as in the origin of hair in mammals or feathers in dinosaurs. Among the different kinds of deviations from recapitulation are heterochrony—changes in timing—and heterotopy—changes in spatial position in a structure. Modified from Sánchez-Villagra (2012), based on Wägele (2005)

Full size image

Comparisons reveal that the limbs are at different stages of development in relation to other structures, highlighting changes in relative timing (Richardson et al. 2009). For example, the forelimbs in marsupials are well-advanced in comparison with many other features, in contrast to any of the other vertebrates (Bininda-Emonds et al. 2007). This is because they have to climb up the fur of the mother to reach her teats short after birth (Werneburg and Spiekman in press).

In amniotes the head is large and distinctly marked off from the trunk, and the heart and the liver form a large bulge and develop early (Richardson et al. 1997). In fishes generally (e.g., zebrafish: Richardson et al. 1997) the heart has not yet formed at stages in which in amniotes the heart has complete looping (Jeffery et al. 2002). This reflects the higher complexity of the amniote heart, which needs more time to differentiate and hence starts to develop earlier (Starck 1979–1982).

There is a clear and simple relation that can be found between some patterns of organogenesis and adult form, related to body elongation and reduction of limbs. In many vertebrates body elongation is accompanied with a larger number of body segments and a reduction of limbs (Müller et al. 2010; Pough et al. 2012; Head and Polly 2015). Among the embryos, one can notice an inverse relationship between somite number and limb bud size (Richardson 1999; Keyte and Smith 2012).

In general, the earlier a structure appears in organogenesis, the larger its size or the greater its complexity in adults, because it has more time to develop (Werneburg et al. 2015). Compared to other mammals, jaw characters in humans occur later, coupled with the fact that our “snouts” are very short compared to other species. The early developmental appearance of our limb related characters corresponds with our elongated limbs as adults.

A video on comparative organogenesis in mammalian evolution

The video, accompanied by a basic audio explanation, portrays prenatal transformations of individuals of different species, embedded in a tree of phylogenetic relationships (Fig. 5). For each species, simple drawings of embryos at different stages were integrated into an animation of transformation. As such, at once, an evolutionary tree depicts not just adults but ontogenies of species. The history of life is a history of life histories.

Fig. 5
figure 5

Snapshots of the video on mammalian organogenesis, available in Supplement 3–10 (eight different languages) of this paper and under following link: https://www.youtube.com/user/SULACOgraphics. a Embryos of three placental mammal species, including humans, are compared in their development. b The phylogenetic arrangement of the depicted species follows Meredith et al. (2011) with modifications following references in Koyabu et al. (2014). c The reconstructed embryogenesis of the last common placental ancestor (Werneburg et al. 2016). d Animation of the hatching of an early amniote, illustrated for Dimetrodon (Synapsida, Permian) as example, with e an adult providing food, illustrating parental care

Full size image

The video first shows the human, rabbit, and deer. The close-up of these three species first familiarizes the viewer with the embryos and makes obvious that there are similarities and differences among them. Each film represents a portion of a process that starts with the fertilized egg and progresses to the completed organogenesis; the animations of transformations (see also Additional files 11, 12, 13, 14, 15, 16, 17) illustrate that embryonic development does not occur in steps or stages, as unavoidably represented on the normal plates, but in a continuous transformation.

Then the tree is shown, in which 21 other species are depicted in their relationships. The study of all these species using parsimony methods serve for the reconstruction of the organogenesis in the last common ancestor of placental mammals (Werneburg et al. 2016). There is then another close-up, that of outgroup representatives, namely a marsupial, a monotreme, and a lizard. Those species are necessary to root the placental tree and to reconstruct the ancestral sequence of character development.

Discussion

The exercise presented here explicitly and implicitly treats different subjects that are fundamental for teaching evolution, tree thinking, and evolutionary mechanisms. In what follows, we discuss some of these subjects and summarize current knowledge on central aspects of mammalian developmental evolution that could be integrated in the teaching on this subject.

Tree thinking. There are many aspects to public communication about evolution, but a fundamental one that would substantially help to correct misconceptions is to associate evolution with evolutionary trees. The presentation of tree-like patterns to depict genealogical relationships among species corrects misconceptions of evolution (Kutschera 2009; MacFadden et al. 2012; Scheyer et al. 2015) and even serves to increase the acceptance of evolution at the university level, according to a study on a population of US American college students (Walter et al. 2013). To provide an effective understanding involves demonstrating macroevolutionary patterns of evolutionary change, as it is the major transitions over long evolutionary time, such as the emergence of limbs in land vertebrates (Laurin 2010), the turtle shell (Scheyer et al. 2013), or the origin of whales (Thewissen 2014), that the layman wishes to understand (Padian 2010; Sánchez-Villagra 2012; Maier and Werneburg 2014).

The general public commonly and wrongly perceives evolution as representing improvement, being progressive and deterministic (discussed by Gould 2002; Zachos 2016). This misconception even reaches the language used by scholars and professional communication on evolution: Rigato and Minelli (2013) studied thousands of publications in the most renowned journals and found in them hundreds of cases of terms and expressions in agreement with the pre-evolutionary metaphor of the scala naturae or the great chain of being [discussed by (Lovejoy 1936; Rieppel 1989)], as when contrasting ‘lower’ to ‘higher’ representatives of a given branch of the tree of life. There is much evidence that even professional biologists lack a true understanding of phylogenetic trees (Morrison 2013). The ‘classic’ linear progression of the ape into the erected human is the most common image to be retrieved in searches for ‘evolution’ on the world wide web. This image is wrong, as the chimpanzees and humans have a common ancestor and both of them have a common ancestor with gorillas and all of them with the orangutan. All apes (incl. humans) are descendants of their last common ancestor.

Divorcing the pattern of common descent from mechanisms and emphasis on macroevolution. The exercise and the video presented here deal with the patterns of morphological changes in development. Leaving aside the mechanisms behind these patterns has many advantages. First of all, it divorces the pattern of common descent from what is generally understood as central to evolution, namely natural selection. The theory of evolution has experienced a significant conceptual and methodological expansion much beyond the Darwin-Wallace theory of natural selection (Gould 2002; Schmid and Bechly 2009; Zrzavý et al. 2013; Laland et al. 2014), yet allusions to the ‘survival of the fittest’ as the sole or most important component in it are rife (Safina 2010). In many texts, evolution is wrongly called ‘Darwinism’ (Scott and Branch 2009), raising an association with ‘social Darwinism’ and the rejection it provokes.

A second positive aspect of the exercise and the video presented here is the macroevolutionary perspective, concerning the large patterns and processes of differentiation at the level of species and above. As convincingly argued by Padian (2010), this aspect in education in evolutionary biology is largely neglected although an understanding of the major evolutionary transitions in the history of life would greatly contribute to diminish uncertainty about evolution (Maier and Werneburg 2014).

Integrating traditional embryology with modern analytical techniques and concepts and a comprehensive study of comparative organogenesis in mammals. The exposure to older works in comparative embryology provides the student with an appreciation of past works that involved careful anatomical documentation. The preponderance of older literature when revising the existing descriptions of development of vertebrates may give the impression that this is an outdated kind of research. This is not the case. Many works have emphasized the necessity to use comparative and quantitative approaches to document the evolution of the phenotype in parallel to experimental and genomic studies, and for that the expansion of the set of model species for developmental studies is fundamental (Jenner and Wills 2007; Milinkovitch and Tzika 2007). Among the recent descriptions of staging systems or developmental series in mammals are those of some bats (Cretekos et al. 2005; Tokita 2006; Wang et al. 2010), tenrecs (Werneburg et al. 2013a), and the echidna (Werneburg and Sánchez-Villagra 2011). These studies have been stimulated not only by the experimental approaches to understand evolutionary novelties arising in development and involving molecular biology (e.g., Sears 2011; Tokita et al. 2012; Montandon et al. 2014). The establishment of quantitative methods to compare developmental timing among species (e.g., Smith 2001; Germain and Laurin 2009; Maxwell and Harrison 2009; Goswami et al. 2016) has also stimulated analyses of accumulated knowledge, revisions of the anatomy of model species previously undocumented (Hautier et al. 2013; Werneburg et al. 2013b), and new studies on the comparative embryology and perinatal life of mammals (Bininda-Emonds et al. 2003), as in our research which was the basis of the animation presented here (Werneburg et al. 2016).

In Werneburg et al. (2016), we integrated information on organogenesis for two monotreme, ten marsupial, 66 placental species (five atlantogenatans and 61 boreoeutherians) and six sauropsids and a lissamphibian. Based on the ‘standard event system’ (SES) of Werneburg (2009), we documented the timing of 123 developmental events, and reconstructed using phylogenetic methods the developmental sequence and timing of organogenesis events in the last common ancestor of placental mammals. The main conclusions of that work are summarized as follows.

There is a mosaic-like pattern of life history traits throughout mammalian evolution. Viviparity evolved in the last common ancestor of marsupials and placentals, the last common ancestor of Theria (marsupials + placentals). The therian ancestor was intermediate between marsupials and placentals concerning altriciality, but the newborn resembled more the ancestral placental anatomy than the marsupial one. Mammals feature diverse levels of maturity at birth, ranging from altriciality to precocity. As is well known, in marsupials the hind limbs are less developed at birth and the forelimbs, used for climbing up the fur of the mother in order to reach the teats (Tyndale-Biscoe 2005), are well-formed. Compared to the therian condition, marsupial gestation length was reduced. The perinatal anatomy of the last common ancestor of placentals differs from that of marsupials. The placental newborn was probably altricial; it probably had closed eyes and an almost naked skin and its limbs were evenly developed. We reconstructed a litter of four young. The developmental innovations in placental mammals include a relatively shorter time until eyelid opening after birth and a longer gestation (125 days) than in the last therian common ancestor.

On the fossil record. The macroevolutionary perspective presented here poses the question on what role paleontology can play in developmental evolution of mammals (Pieretti et al. 2015). The evolutionary history of the synapsid lineage since the divergence from the sauropsid (reptiles and birds) sister-group in the Carboniferous (Benton et al. 2015) is documented by a growing fossil record that documents the tempo and mode of acquisition of the many diagnostic features of Mammalia (Angielczyk 2009). The fossil record also documents features that reveal changes in growth patterns and markers of life history such as dental replacement (Sánchez-Villagra 2010; O’Meara and Asher 2016), but the direct record of organogenesis is almost non-existent (Franzen et al. 2015). As such, the fossil record is mute on the subject treated here, but it does provide the evolutionary time in which the groups in question diverged.

Conclusions

The proposed activity addresses the evidence-based process of science among the core competencies of the Vision and Change report on education in undergraduate biology (Brewer and Smith 2011). The developmental patterning of the ‘body plan’ of animals is determined by complex and multi-genic interactions (Held Jr. 2014). The role of Hox genes and other genes in this process is usually the subject of courses, whereas the phenotypic transformations that occur in the individual development and the changes on such transformations in geological time are in many cases neglected. The activity presented here serves to address this deficit with an effective exercise that combines concepts of development and evolution. The use of a developmental perspective can bring great insights into teaching human anatomy even from a clinical perspective (Diogo et al. 2016).