To the Editor:Stuart Newman

In considering scenarios for the evolution of pattern in the

vertebrate limb, Cohn et al.(1) make a strong case that branch-

ing of skeletal elements by bifurcation at the tips of existing

elements is not an accurate representation of the develop-

mental process. From this they conclude two things. The first,

whichseemsreasonable, is thata ‘physical-mechanical’model

that entails such branching(2,3) provides a poor account of both

ontogeny and phylogeny. But they then go on to create the

misleading impression that all accounts of skeletal develop-

ment based on the physical properties and self-organizing

capabilities of precartilage mesenchyme are equivalently uni-

formative with respect to evolutionary transitions.

The gist of their argument is as follows: connections be-

tween skeletal elements in extant or extinct species that

appear to represent branches of other elements have been

used to map pathways of phylogenetic descent. But if there is

no actual embryonic branching process, the observed con-

nections have no logic beyond the ‘details of developmental

and genetic regulation of patterning andmorphogenesis.’ This

latter set of processes they identify with ‘positional informa-

tion,’ in which cell and skeletal element identity is presumed

to be specified by local values of a molecular grid(4) and the

physical process by which condensations form has no bearing

on the origination of pattern elements.

A different physical model for mesenchymal morphogen-

esis(5) is based on Turing’s ‘reaction–diffusion’ scheme for

producing standingwavepatternsof chemical concentration(6)

rather than on mechanical instabilities in a compressible

mediumaswas themodel(2) thatwas the object of Cohnet al.’s

critique. The reaction–diffusion model makes use of standard

properties of developing tissues—the secretion of diffusible

growth factors that activate and inhibit differentiation, in this

case, chondrogenesis. With the discovery that the secreted

activator of mesenchymal matrix production, TGF-beta, is

positively autoregulatory(7) a class of candidates emerged for

the putative Turing morphogen in the limb—a molecule that

could participate in a self-organizing pattern-forming network

and, at its spatial peaks, induce precartilage condensations by

adhesive interactions.(8,9) Ectopic administration of TGF-beta

in the interdigitial regions of the developing autopod produces

extra digits.(10)

Recent experiments indicate that TGF-beta2 indeed acts

as a Turing morphogen in setting the pattern of precartilage

condensations in vitro.(11) Most importantly for the issue at

hand, experiments designed to test competing models for

mesenchymal pattern formation led to results consistent with

the reaction–diffusion/differential adhesion model, but incon-

sistent with the mechanical instabilty model.(12)

Examination of Turing-type reaction–diffusion processes

(a range of simulation results and animations of one such

system can be found at www.cacr.caltech.edu/ismap/image.

html) shows that small parameter changes (which would

correspond to altered rates of biosynthesis, diffusion, and so

forth in a developing tissue) can lead to the emergence of

novel pattern elements (nodular or rod-like condensations, for

example), without the necessity of bifurcation or branching

from existing elements.

Cohn et al. correctly point to the fact that the de novo

emergence of a new precartilage condensation, and thus

skeletal element, during the course of evolution does not carry

with it an implication as to its relation to existing elements. But

origination and autonomization of structural elements are

distinct phases of morphological evolution, both of which need

to be considered in phylogenetic scenarios.(13) Focusing

solely on the molecular processes underlying the realization

of structures inmodern-day organisms, as Cohn et al. seem to

suggest, risks missing the interplay between self-organizing

tissue properties and the reinforcing genetic mechanisms that

mayarise later and stampaphysically originated structurewith

an individual identity.(14,15)

References1. Cohn MJ, Lovejoy CO, Wolpert L, Coates MI. Branching, segmentation

and the metapterygial axis: pattern versus process in the vertebrate limb.

BioEssays 2002;24:460–465.

2. Oster GF, Murray JD, Harris AK. Mechanical aspects of mesenchymal

morphogenesis. J Embryol Exp Morphol 1983;78:83–125.

3. Shubin NH, Alberch P. A morphogenetic approach to the origin and

basic organisation of the tetrapod limb. Evol Biol 1986;1:319–387.

4. Tickle C, Summerbell D, Wolpert L. Positional signalling and specification

of digits in chick limb morphogenesis. Nature 1975;254:199–202.

5. Newman SA, Frisch HL. Dynamics of skeletal pattern formation in devel-

oping chick limb. Science 1979;205:662–668.

6. Turing A. The chemical basis of morphogenesis. Phil Trans Roy Soc

Lond B 1952;237:37–72.

7. Van Obberghen-Schilling E, Roche NS, Flanders KC, Sporn MB, Roberts

A. Transforming growth factor beta-1 positively regulates its own ex-

pression in normal and transformed cells. J Biol Chem 1988;263:7741–

7746.

8. Newman SA, Frisch HL, Percus JK. On the stationary state analysis of

reaction-diffusion mechanisms for biological pattern formation. J Theor

Biol 1988;134:183–197.

9. Newman SA. Lineage and pattern in the developing vertebrate limb.

Trends Genet 1988;4:329–332.

10. Merino R, Ganan Y, Macias D, Economides AN, Sampath KT, Hurle

JM. Morphogenesis of digits in the avian limb is controlled by FGFs,

TGFbetas, and noggin through BMP signaling. Dev Biol 1998;200:

35–45.

BioEssays 24:1077–1078, � 2002 Wiley Periodicals, Inc. BioEssays 24.11 1077

Correspondence

11. Miura T, Shiota K. TGFbeta2 acts as an ‘activator’ molecule in reaction-

diffusion model and is involved in cell sorting phenomenon in mouse limb

micromass culture. Dev Dyn 2000;217:241–249.

12. Miura T, Shiota K. Extracellular matrix environment influences chondro-

genic pattern formation in limb bud micromass culture: Experimental

verification of theoretical models. Anat Rec 2000;258:100–107.

13. Muller GB, Newman SA. Generation, integration, autonomy: three steps

in the evolution of homology. Novartis Found Symp 1999;222:65–73.

14. Salazar-Ciudad I, Newman SA, Sole R. Phenotypic and dynamical transi-

tions in model genetic networks. I. Emergence of patterns and genotype-

phenotype relationships. Evol Dev 2001;3:84–94.

15. Salazar-Ciudad I, Sole R, Newman SA. Phenotypic and dynamical

transitions in model genetic networks. II. Application to the evolution of

segmentation mechanisms. Evol Dev 2001;3:95–103.

Stuart Newman

E-mail: NEWMAN@NYMC.EDU

DOI 10.1002/bies.10158

Published online in Wiley InterScience

(www.interscience.wiley.com).

Correspondence

1078 BioEssays 24.11