4
Excitable Media in Medias Res:
How Physics Scaffolds Metazoan Development and Evolution
Stuart A. Newman
A naive notion of morphological evolution holds that organismal form can change in arbitrary ways under selective regimes. Experience tells us, on the contrary, that different classes of multicellular organisms—animals, fungi, plants—exhibit morphological motifs that (despite some ambiguities) stamp their members with the group’s identity. What are the underlying generative principles that produce these forms, ensure that they are inherited in a type-specific fashion, but also allow for the possibility of evolutionary transformations? A solution to this problem will inevitably involve the concept of scaffolding. The perpetuation over time of a set of structures, where the primary elements themselves are not carried forward, but the means by which they are constructed are, implies the existence of an implicit organizational framework.
Research over the past two decades has shown that with respect to the multicellular animals, the products of subsets of a common set of ancient genes—the “developmental toolkit”—which first evolved in single-celled ancestors (King et al. 2008; Shalchian-Tabrizi et al. 2008), are employed for the generation of pattern and form at the embryonic stages of each phylum’s members (Wilkins 2002; Carroll et al. 2004). This toolkit contains transcription factors, which (along with their cognate cis-acting regulatory elements) determine cell-type identity, as well as molecules involved in cell–cell interaction (e.g., adhesion and communication). The latter molecules function in the three-dimensional arrangement of cells to produce bodies and organs (Newman et al. 2009).
However, the gene products constituting the “cell interaction toolkit” do not act on their own to generate organismal form. More specifically, embryogenesis is not simply a matter of self-assembly of proteins and other molecules. Living tissues, including the primordial cell clusters that founded the first metazoan lineages, are physical materials, and many of their morphological states and transformations can be understood in terms of “generic” properties they share with nonliving viscoelastic materials, particularly those (“excitable media”) that store mechanical and chemical energy (Newman and Comper 1990; Forgacs and Newman 2005). We have proposed that the main role of the interaction toolkit is to mobilize physical forces and processes of the “middle scale” (mesoscale: 100 μm–10 mm) relevant to the dimensions of multicellular aggregates (Newman and Bhat 2008, 2009; Newman 2010, 2012).
In conjunction with this hypothesis we have coined the term “dynamical patterning modules” (DPMs) for the functional units composed of one or more interaction-toolkit molecules and the physical effect they mobilize. In most cases (as discussed in the next section) the multicellular functions of the DPM-associated molecules are emergent properties of their coming to operate on the multicellular scale. The incorporation of mesoscale physics into the explanation of morphogenesis thus provides a basis for understanding the emergence of complex animal forms in what appear to have been two time-compressed episodes in the late Precambrian and early Cambrian periods (around 600 to 540 mya) (Rokas et al. 2005; Shen et al. 2008).
In this view, the physics of mesoscale materials provides a scaffold for the organization of animal form (Forgacs and Newman 2005; Newman et al. 2006). This is distinct from the claim that animal embryos are merely mesoscale materials. Genetics also plays a decisive role: only the expression of certain genes permits cell clusters to mobilize particular physical forces on this scale. The specificity inherent in the presence or absence of given genes is therefore reflected in specificity at the physical level (Newman 2011a). How the interaction-toolkit products harness various physical processes is the subject of earlier publications (Newman and Bhat 2008, 2009) and will be summarized in the following section. This perspective will then be applied to the solution of a long-standing puzzle in evolutionary developmental biology, the so-called “embryonic hourglass.” In the remaining portion of the present section I will briefly describe the scaffolding role of mesoscale physics in development and evolution.
According to Caporael, Griesemer, and Wimsatt (“Developing Scaffolds: An Introduction,” this volume), scaffolding is “(1) facilitation of a process that would otherwise be more difficult or costly without it, which (2) tends to be temporary—an element of the maintenance, growth, development, or construction process that fades away, is removed, or becomes ‘invisible’ even if it remains structurally integral to the product.” In the DPM-based framework for the origination, generation, and evolutionary transformation of biological form, mesoscale physics is indeed a necessary but transient condition for all these phenomena. The morphological motifs generated by the DPMs will persist even as the physico-genetic bases for their first emergence wane or even vanish.
The DPMs provide a material basis (with respect to morphological development and evolution) for three integrating perspectives that have jointly motivated the analysis of complex natural systems in terms of the scaffolding concept (Caporael, Griesemer, and Wimsatt, introduction to this volume). First, it exemplifies the reproducer perspective (Griesemer 2000a,b; 2005) in its recognition that forms are produced anew in each generation not simply by the action of genes but by the unified activity of gene products and other organizing principles—in this case the physical processes mobilized by the products of those genes. Matter of a given composition and scale will “inherit” the physical laws that pertain to it just as (if it is also living matter) it inherits its genes.
The DPM framework also provides an analogue, at the level of “cell societies,” to what has been termed (with respect to human social organizations) the core configurations model (Caporael 1995, 1997; Caporael and Barron 1997). This model describes specific emergent properties that characterize interactions first between two, and then among successively larger numbers of individuals. As will be seen in the following section, certain physical effects that have little impact on entities of the scale and composition of single cells have organizational effects on pairs of cells and on multicellular clusters.
Finally, DPMs will be seen to shed light on a property of developmental systems termed generative entrenchment (Wimsatt and Schank 2004; Wimsatt 2001). In such systems certain processes or configurations are integral to the final outcome in the sense that if they are compromised, numerous downstream events will fail. If the system evolves, it will not readily tolerate changes to these entrenched elements.
The most entrenched steps or outcomes in a developmental process are often the earliest ones, but not always (Rasmussen 1987). Many embryos, for example, show regulative behavior in which they can recover the capacity to produce a normal result after early-stage disruptions but not later ones. Moreover, despite the expectations of a simply mechanistic concept of developmental programs, the developmental steps acquired earlier in the course of evolution are not necessarily more generatively entrenched than later-evolved ones. The mobilization of mesoscale physics partway into the developmental process, discussed in detail below, helps explain the midstream entrenchment of what appears to be (but as I will show, is not precisely) the “phylotypic stage,” the stage at which the phylum-specific morphological motifs or gene expression patterns first appear (Sander 1983; Slack et al. 1993), a phenomenon of comparative embryology known as the embryonic hourglass.
Functions and Components of the Dynamical Patterning Modules
A variety of basic functions, dependent on the ability of ancient molecules (the cell interaction toolkit mentioned above) to mobilize physical forces and effects, drive pattern formation and morphogenesis in present-day embryos and are likely to have served analogous functions at the origin of these processes (reviewed in Forgacs and Newman 2005 and Newman and Forgacs 2009). While some choanozoans (the closest modern unicellular relatives of the Metazoa) achieve multicellularity by retaining cytoplasmic bridges after division (Dayel et al. 2011), some members of this group also contain genes that specify cadherins (Abedin and King 2008), which mediate cell–cell attachment in metazoans. It is plausible that metazoan multicellularity arose by the recruitment and repurposing of such molecules for aggregation, but if so, the instability entailed by the potential genomic nonuniformity of the cells of the resulting clusters (Grosberg and Strathmann 2007) would have presented a new evolutionary challenge. I have suggested that this was the impetus for the evolution of an egg stage of development, which provided a means to regenerate, by cleavage, the ancestral cell cluster, but now containing cells with identical genomes (Newman 2011b).
Once multicellularity was achieved in proto-metazoans, the biological entities in question were no longer the internally highly structured parcels of matter surrounded by inextensible membranes that individual cells are. In a multicellular aggregate the cells assume the role of loosely associated and independently mobile subunits of what is, in a formal sense, a liquid droplet. Liquids exhibit surface tension, and if their subunits do not exhibit any marked anisotropy, their droplets will contain no internal spaces and be spherical by default (due to the principle of energy minimization). Similarly for an aggregate of cells: if its cellular subunits are uniform in their surface adhesive properties, any interior spaces will automatically become filled in and, if cell shape is relatively isotropic, the cluster’s shape will be a sphere. If, however, the individual cells become polarized in either surface composition or shape, energy minimization will drive such aggregates to become hollow or elongated.
Another property liquid-like tissues have in common with nonliving liquids is the capacity to “phase separate” when there are different subunits with different affinities for each other. Energy minimization, for example, leads to tissue multilayering if cells reliably differ (qualitatively or quantitatively) with respect to the expression of adhesion molecules (Steinberg 2007).
All cells potentially exhibit multistable dynamics due to the properties of their internal gene regulatory networks (reviewed in Forgacs and Newman 2005). The physical effect that enables a multicellular cluster to maintain a balance of different cell types is lateral inhibition, whereby a cell signals adjoining or nearby ones to assume a different state than its own (Meinhardt and Gierer 2000). In addition, the secretion of mobile molecules (morphogens) transported by diffusion or related processes, permits the multicellular aggregate to develop chemical gradients, making it different from one end to the other, an effect that promotes spatially dependent cell differentiation (Lander 2007).
Oscillation in internal chemical composition, a behavior potentially sustained at multiple biochemical levels by any cell (Reinke and Gatfield 2006), has the reciprocal effect since the oscillations spontaneously and inevitably come into synchrony at the multicellular level (Strogatz 2003; Garcia-Ojalvo et al. 2004), generating long-range coordination of cell state, that is, morphogenetic fields (Gilbert 2010). When synchronized oscillations or related circuitry (e.g., the reaction-diffusion mechanism of Turing 1952, reviewed in Kondo and Miura 2010, or the local autoactivatory–lateral inhibitory mechanism of Meinhardt and Gierer 2000) interacts with morphogen gradients, the result may be the periodic or quasi-periodic arrangement structures, such as the skeletal elements of the vertebrate backbone (Lewis et al. 2009) or the paired limbs (Zhu et al. 2010).
Lastly (for this discussion, although several additional patterning functions have been described; see Salazar-Ciudad et al. 2003; Newman and Bhat 2009) are extracellular matrices (ECMs), which can cause epithelial cell sheets to resist bending deformations and promote invagination, evagination, and branching morphogenesis, or turn liquid-like mesenchymal aggregates solid.
Although a characteristic spatial scale and set of material properties (i.e., isotropic or anisotropic subunits, viscoelasticity, chemical and mechanical excitability) are required to enable the DPMs, the molecules and pathways that mediate all the above-mentioned processes in extant embryos were already present either in full or in major part in single-celled ancestors (reviewed in Newman and Bhat 2008, 2009; Newman et al. 2009). This is evidenced by the occurrence of corresponding sequences in the genomes of choanozoans, or fungi, whose common ancestor with the animals diverged even earlier. Examples include cell–cell and cell–ECM adhesion molecules (e.g., cadherins, lectins, integrins), ECM molecules themselves (e.g., collagens), intracellular components of the Wnt pathways, which act at the cytoskeletal level to control cell surface and cell shape polarization, receptors and ligands, as well as nuclear mediators, of the Notch pathway (on which multicellular lateral inhibition depends), and morphogens, such as Hedgehog and bone morphogenetic proteins (Mendoza et al. 2005; Hurov and Piwnica-Worms 2007; Abedin and King 2008; King et al. 2008; Shalchian-Tabrizi et al. 2008; Exposito et al. 2010; Sebé-Pedrós et al. 2010).
Dynamical Patterning Modules and the “Morphogenetic Stage”
Dynamical patterning modules, as we have seen, came into existence not by the evolution of new genes but by virtue of the capacity of the ancient gene products described above to mobilize mesoscale physical forces, processes, and effects when they came to operate in cell clusters. Significantly, all present-day embryos pass through an ontogenetic stage analogous to the phylogenetic stage of primitive cell clusters in which DPMs first appeared. Corresponding to (depending on the species), the morula, blastula, blastoderm, or inner cell mass, this “morphogenetic stage” (see also Seidel 1960), consists of dozens to scores of identically sized cells and is the end product of cleavage of the fertilized egg. It is also precisely the stage of development at which the embryo becomes a parcel of mesoscopic matter and all the DPMs characteristic of the embryo’s phylum are operative.
The morphogenetic stage occurs earlier and consists of fewer cells than what has previously been described as the “phylotypic stage,” but I suggest that it is a more helpful concept for thinking about the relationship between development and evolution, and the role of mesoscale physics in scaffolding both these processes.
Premetazoan cell clusters (those that formed by aggregation as well as those that formed by the subdivision of an enlarged cell, or “proto-egg”; Newman 2011b) would have expressed DPM-associated molecules characteristic of the genomes of their single-celled progenitors. Since some of these molecules were shared and some unique, these clusters would have embodied partly overlapping sets of DPMs. Specific complements of DPMs, in fact, may be the most useful causal characterization of the various animal phyla (Newman et al. 2009; Newman 2011b, 2012). The earliest radiating metazoan phyla, the two-layered diploblasts, lack one or another of the basic DPMs (Srivastava et al. 2008, 2010), but the later-appearing three-layered triploblasts appear to contain all of the DPMs. The triploblastic phyla, whose member organisms contain connective tissues, can be distinguished from one another by the molecular and physical natures of their ECMs.
The DPMs act combinatorially to generate a circumscribed set of morphological motifs. For example, an organism that maintained itself in distinct, ordered layers would need to have polarized cells and therefore to contain and express genes of the Wnt pathway (which appears to function uniquely in cell polarization in the metazoans.) The single known placozoan species, Trichoplax adhaerens, has its four cell types arranged into homogeneous layers, except for the basal layer which contains two of the types. It indeed expresses the canonical (apicobasal polarity-determining) branch of the Wnt pathway (Srivastava et al. 2008). True animal embryos organize cells into complex tissues and patterned arrangements of cells utilizing the Notch pathway. Trichoplax lacks these histological features and also the Notch ligands and receptors (Srivastava et al. 2008). Sponges are another group of diploblastic forms, but ones which do not maintain strict boundaries between epithelia-like and mesenchymal-like tissues. Correspondingly, they lack key ECM components of the basal lamina (Srivastava et al. 2010).
While the morphological motifs characteristic of the embryos of a given type of organism will thus be a function of the DPMs that operate during its development, there are many variations on each phyletic theme. These variations are associated with subphyla and classes within the phyla (e.g., fish, amphibians, reptiles/birds, and mammals, within the vertebrate subphylum of Chordata), and may have provided the raw material for their adaptive diversification.
Because the DPM-capacitating morphogenetic stage occurs partway into development, it is not immediately obvious that it would be the stage at which the scaffolding function referred to above is implemented and (as I will show below) the most generatively entrenched portion of ontogenesis. In fact, embryonic development is unexpectedly tolerant to intracellular variations at the postfertilization stages of development preceding the morphogenetic stage.
Egg-Patterning Processes: Extensive Variation in Ooplasmic Organization with Conservation of Phylotype
In this section I describe a class of processes that like DPMs are based in part on physical effects but, unlike the latter, do not act at the multicellular level and therefore do not scaffold the development and evolution of form. Acting earlier than the DPMs, within the confines of the individual egg cell, this separate set of processes, the “egg-patterning processes” (EPPs), modulate and fine-tune the scaffolding role of the latter but, as I will show, cannot push development beyond the bounds of the DPM-determined phylotypic outcomes.
One important way that variability of form within the various phyla is manifested and implemented is by evolutionary alterations in egg size and shape, and most specifically, in the intra-egg patterning processes that occur prefertilization (i.e., during oogenesis) and postfertilization. Comparing the eggs of the different vertebrate classes—the huge, yolky bird egg, the microscopic eutherian mammal egg, the intermediate-sized egg of fish, amphibians, and nonavian reptiles—shows the lack of obvious mapping between egg morphology and body plan (which is similar in all these examples).
Oocytes and pre- and postfertilized eggs are rendered internally nonuniform by two kinds of processes. The first mode, common to all animal taxa except for a few (e.g., eutherian mammals), employs cytoplasmic determinants (“ooplasms”) incorporated into distinct regions of the egg during oogenesis. The EPPs are physical and physicochemical effects induced by sperm entry or parthenogenetic activation (Newman 2009, 2011b). The cytoplasmic heterogeneities generated by either or both modes, though often associated with recognizable polarities and landmarks of the adult stage, do not correspond to maps or blueprints of the subsequently developed organism (see, e.g., Freeman 1999).
EPPs are based on single-cell cytoplasmic functionalities that evolved before the existence of eggs, and indeed of multicellularity (Newman 2009, 2011b). Spatial domains of the dimensions of a single cell (~10 µm) would not typically sustain mesoscale physical processes. However, the expansion of cell size by an order of magnitude or more during the evolution of the egg stage of development would have transformed cell physiological functions that had evolved to operate on the microscale into mesoscale effects (Newman 2011b). In almost all cases, the relevant intracellular processes have some “generic” physical aspects, in that they are based on material properties and capabilities such as diffusion, viscoelasticity, sedimentation and convection, and chemical excitability, which are common to living and some nonliving systems (Newman and Comper 1990). In this sense they are analogous to the DPMs. However, the two sets of processes differ in a key property: whereas DPMs do not exist apart from the multicellular state, the EPPs are entirely intracellular. Since the generic physical components of the EPPs have inherent propensities to organize matter in preferred directions, sometimes in abrupt, nonlinear ways in response to changes in system parameters (see below), the degree to which they can be incrementally molded by natural selection is limited.
The DPMs, by definition, can organize the cell aggregates of the morphogenetic stage of embryogenesis without the need for any prespecification in the egg, acting on clusters of identical cells generated by simple cell multiplication, or by cleavage of an internally homogeneous egg. When an egg is rendered internally heterogeneous by the action of one or more EPPs, its cleavage leads to a cluster of cells that differ from one another across the cluster in a continuous or discontinuous fashion. The morphogenetic stage embryo is then said to exhibit axial polarity.
Generating such polarities provides advantages for the propagation of a biological type from one generation to the next. A “reproducer” (Griesemer 2005) that passes through an egg stage that exhibits polarity will present a set of “initial conditions” to the DPMs mobilized at its morphogenetic stage. Although these DPMs would be the same whether or not the system had been “prepared” in this fashion, a well-known property of dynamical systems is that their outcomes are more reliable when their initial conditions are preset rather than random (Sokolnikoff and Redheffer 1966). This being said, the phylotypic morphology arising from an egg in which EPPs were operative would still be the same as one which had arisen from a homogenous egg, or from one in which a different set of EPPs had operated, since the available DPMs would be no different.
In a widely discussed example, long and short germ band insects (e.g., fruit flies and beetles), exhibit vast differences in pregastrulation development (including intracellular organizational processes) before arriving at similar body plans, illustrating the insensitivity of morphogenetic-stage events to the preparation of the system to that point (reviewed in Salazar-Ciudad et al. 2001). Here I will illustrate the concept of EPPs and the relative inconsequentiality of variations in their action with respect to body plan organization using the less familiar example of the comparative developmental biology of nematode worms.
The nematodes are probably the most diverse of all the animal phyla, being represented by up to 1,000,000 species. Apart from their size, however, which can range in body length from a few millimeters to a few centimeters, the anatomy of these worms is nearly indistinguishable. In spite of this morphological conservation, postfertilization events, and specifically the deployment of EPPs, differ dramatically within this group.
The egg of the nematode C. elegans is unpolarized before it is fertilized. Upon sperm entry the egg’s cortical cytoplasm becomes reorganized, resulting in an asymmetrical distribution of various factors before the first cleavage. This polarity is required for the establishment of the anteroposterior (A-P) axis during embryogenesis (Rohrschneider and Nance 2009). The reorganization of the egg’s interior involves cortical cytoplasmic flows that depend on both the contractile protein complex actomyosin and the activity of sperm-contributed microtubules. These flows cause the enrichment in the anterior region and depletion in the posterior region of the egg of an initially uniformly distributed enzyme complex. This process, in turn, causes a different enzyme complex to accumulate in the posterior region. As a result, although the sperm does not attach at a preferred site on the egg, its entry point defines the future posterior pole (Munro et al. 2004; Munro 2007; Tsai and Ahringer 2007).
Irrespective of the elaborate fashion by which A-P polarity in C. elegans is established, the means by which A-P symmetry is broken, and even the developmental stage at which it occurs, can be extremely different in other nematodes whose final forms are essentially identical.
In the nematode B. xylophilus, for example, which is anatomically indistinguishable from C. elegans, the sperm entry point becomes the future anterior pole of the embryo and the pattern of cortical flow and its relation to the sperm microtubules are entirely different from that in C. elegans (Hasegawa et al. 2004). In another worm, R. culicivorax, the first cleavage is symmetrical rather than asymmetrical, and the pattern of subsequent asymmetrical cleavages and alternative assignment of cell fates suggests that A-P polarity is determined in still a different fashion from the other two nematode species (Schulze and Schierenberg 2008).
In the freshwater nematode T. diversipapillatus, no asymmetrical cleavages and no distinct cell lineages are generated until the morphogenetic stage, which resembles that of all nematodes previously studied, but rather than starting as a solid ball of cells begins as a hollow blastula (Schierenberg 2005). And, among many more possible examples, in three different parthenogenetic species of nematodes, which have no opportunity for sperm entry to influence the assignment of A-P polarity, this feature is acquired in ways that differ from the other instances mentioned, and also from one another (Lahl et al. 2006).
Thus, whereas acquisition of A-P polarity is clearly an essential aspect of nematode anatomy, the way that it is acquired during development seems to have little impact on the final morphological outcome, which is always essentially the same.
A consistent and parsimonious explanation for the wide variation tolerated by the nematodes (and other phyla, notably ascidians and vertebrates; see Newman 2011b for a review) in the deployment and outcomes of the EPPs is that postfertilization ooplasmic reorganization, however integral it may become to the derived features of a phylum’s subtaxa, do not set or influence the defining morphological motifs of the phylum. This role is performed by the DPMs, the set of physico-molecular events specifically mobilized at the morphogenetic stage.
Conclusion: The Scaffolding of Development by Mesoscale Physics Finds Confirmation in the “Embryonic Hourglass”
I have described how the wide diversity of form in early embryos within given animal phyla (nematodes being the main example, but mollusks, annelids, arthropods, and chordates providing equally good ones), can be rationalized by an understanding of the roles of dynamical patterning modules in the generation of body plans. The DPMs collectively serve as a scaffold in being both facilitating and transient (Caporael, Griesemer, and Wimsatt, introduction to this volume). In addition, as with many architectural scaffolds, while being modular and therefore partly general purpose, DPM-based scaffolding is also capable of imparting specificity since each phylum is genetically constrained to employ a partly distinctive set of DPMs.
The breadth of pre-morphogenetic-stage variation consistent with stability of phylotypic identity has a counterpart in the variety of fully developed forms that may ultimately be produced within the confines of a given phylotype. Post-morphogenetic stage diversification results from the fact that the DPMs can be employed in different combinations and temporal orders and will manifest differently depending on how the morphogenetic-stage cell aggregates are prepatterned (the latter resulting from both maternal effects and the EPPs). The “variational properties” of the developmental mechanisms available to the species of a given phylum (as reflected in the comparative anatomy of the adult forms of the phylum’s subtaxa) can thus differ extensively from one another (Salazar-Ciudad et al. 2003; Salazar-Ciudad 2010).
Notwithstanding this intraphylum variation however, embryos of any species within a phylum are subject to their group’s particular ways of setting up the initial conditions of the morphogenetic stage. Entrenchment of these “frozen accidents” over the course of evolution will prohibit significant developmental variation within a genus or species.
The passage of all of a taxonomic group’s (e.g., a phylum’s) embryos through a morphologically conserved intermediate stage of development before they go on to assume their subgroup-specific characteristics has been termed the embryonic hourglass (Duboule 1994; Raff 1996; Hall 1997; see also Horder 2008). Were the reproducers (Griesemer 2000a,b; 2005) in these evolutionary scenarios appropriately defined solely in terms of genes and genetic programs, it would make sense to focus on that stage at which the pattern of gene expressions takes a phylum-specific turn (the phylotypic stage) and ask, as many researchers have done, what in the history of these groups led to the stability (and apparent entrenchment) of particular intermediate stages of development despite the ooplasmic variation before and morphological after such stages. For example, is the conserved intermediate stage less susceptible to variation than earlier or later ones because of strong stabilizing selection (Galis et al. 2002)? Does it represent the effect of developmental (Cruickshank and Wade 2008) or functional constraints (Wray and Strathmann 2002) or robustness of developmental mechanisms (von Dassow and Munro 1999)? Or is it just an artifact of subjective criteria in the characterization of form (Bininda-Emonds et al. 2003)?
In the view put forward here, in contrast, the reproducers—multicellular organisms that undergo development—are propagated from generation to generation due not simply to the genes they contain but to the mesoscale physical processes that are mobilized by a subset of those genes once the organism reaches its morphogenetic stage. While the capacity to harness a specific set of physical process is indeed dependent on a particular set of DPM-associated toolkit genes (virtually all of which already existed in the unicellular antecedents of the metazoans), there is no privileged phylotypic stage during an organism’s ontogeny. The reason why the pan-taxonomic distribution of embryonic forms within a phylum takes the form of an hourglass is because all developmental trajectories inevitably pass through a common stage consisting of an aggregated population of similarly sized cells (see figure 4.1).
The DPMs can only mobilize the relevant physical effects in medias res, at the morphogenetic stage. Like the social behaviors considered in the core configurations model of Caporael and coworkers (Caporael 1995, 1997; Caporael and Barron 1997), then, the processes that organize the phylotypic anatomies are emergent properties of a critical number of actors, cells in this case.
The notion of mesoscale physics as a scaffold for the action of genes that mediate cell–cell interactions, that is, the DPM framework, accounts for both the origination and entrenchment of animal form over the course of evolution (Müller and Newman 2005; Wimsatt 2001, 2007). In this perspective, the major morphological motifs of animal body plans arose early, by mechanisms that yielded outcomes that were simultaneously plastic and stereotypical. This led to the explosion of body types seen at the Precambrian–Cambrian boundary more than half a billion years ago (Conway Morris 2006; Shen et al. 2008). Subsequent genetic change, though incapable of taking the developmental trajectories of these multicellular systems outside of the morphospace defined by the DPM-based scaffolding, integrated and consolidated the generation of their forms, ultimately producing the canalized developmental programs of present-day animals.
References
