Abstract 07 Haeckel’s recapitulation of phylogeny in ontogeny has been taken dead in contemporary biology since 1977 due to the lack of its explanation.

Abstract 07

Haeckel’s recapitulation of phylogeny in ontogeny has been taken dead in contemporary biology since 1977 due to the lack of its explanation. Biologists were waiting for a long time for any explanation of it since the fall down of the first explanation proposed by Haeckel – inheritance of acquired characters, destroyed by genetics. Depriving recapitulation of the first explanation was taken as undermining of statistical estimations which lay in the background of recapitulation formulation. Such estimations are hard to prove and stay subjective, therefore they are hard to defend. In the effect, false view has been established that recapitulation cannot be in harmony with genetics and it must be reconstructed to e.g. repetitions.The main postulate of genetics is that changes should appear evenly in whole ontogeny, not only in the end as the recapitulation needs. De Beer uses such an assumption in his pure and - from contemporary point of view – naive model, where ontogeny is only a chain of transformations and adaptive condition describing Darwinian evolution is not used. This model explains terminal modifications regularity and repetition, it is old and long time not discussed but still the latest and up till now in force.The same main postulate of genetics is used in our model but we consider the ontogeny as complex network of conditional transformations randomly evolving under adaptive condition. This is enough to obtain statistical effects named the structural tendencies, equivalent of terminal modifications and terminal addition regularities. Terminal additions tendency is especially important, because it creates recapitulation and lack of knowledge of its creating mechanism was the cause of death of recapitulation. Model’s form of this tendency is the terminal predominance of addition over removal. It is accomplished by tendency of simplifying of deeper part of evolving system, which means: predominance of removing over addition in the earlier area but this tendency does not create the recapitulation – it only makes results of simulation alike Weismann’s ‘pressing back’ conception. We use a wide range of network parameters and types in the simulation experiments. There are Kauffman networks and similar, random Erdős-Rényi networks and nowadays popular, scale-free Barabási-Albert networks and others. These tendencies are present in all of them. We also observe similarity of functional and historical order in our simulations, it is the main element of recapitulation idea. Computer simulation is the powerful tool, which allows to investigate mechanisms of these tendencies in detail, therefore now they are known and recapitulation can come back to biology fully alive.Reanimation of Haeckel’s recapitulation removes one of the biggest blockade in the biology. Since the loss of its first explanation whole discipline - comparative embryology lost the main evolutionary aspect. To reanimate the union of developmental and evolutionary biology, Wilkins has distanced himself from Haeckel in public, now he does not have do it any more. Now recapitulation effects are expected in conditional, complex specification characteristic of all vertebrates and a few invertebrates.

Old Recapitulation of Phylogeny in Ontogeny is Still Alive, It got New Explanation in Complex Network Area Andrzej Gecow

Andrzej Gecow

Institute of Paleobiology Polish Academy of Science,

Warsaw Poland,

[email protected]

11th Evolutionary Biology Meeting at Marseilles 2007

Old Recapitulation of Phylogeny in Ontogeny is Still Alive,

It got New Explanation in Complex Network Area

Evolutionary biology concepts and Knowledge

This investigation is partially founded by Polish Government

Recapitulation of phylogeny in ontogeny

Ontogeny

Phylogenyin the sense: sequence of adult forms

One cellancestor

embryo with gill

geological time

fish

Phenotype

zygote

Fossils

Biogenetic law:( Haeckel 1866)

Parallelism of 3: ( Haeckel 1905)

Mutations

???

ontogenystages(time indays)

It is now negated ...

Parallelism exists or not?

Alternatives:1. recapitulations – ontogenetic stages are similar to adult ancestors2. repetitions – similarities in ontogeny are its stage relics only

2.Change initiation occurs with equal probability in all length of ontogeny, but it causes changes in all later parts of ontogeny. (de Beer 1940)It explains Terminal Modifications and Conservation of early stages regularity. (Naef 1917)- ontogeny change types include additions and removals similarly,- ontogeny changes occur more frequently near the end of ontogeny.

The choice depends on mechanism of ontogeny changes:1. Haeckel suggests: inheritance of acquired characters

but genetics... It gives Terminal Additions regularity. ( „Pressing back”

conception: Terminal Additions and Compression of

early stages. Weismann 1902), - ontogeny expands by additions of new adaptive transformations at its end.

PhylogenyOntogeny

repetitions

ontogenytime

changeinitiations

Lack of a mechanism. Observations must agree with genetics!

After 1. Observations are statistical (indicated by Haeckel) a. and subjective. They must agree with genetics! => recapitulation does not exist! => ‘Terminal Modifications’ is better and replaces ‘false’ ‘Terminal Additions’ => They are difficult to defend. Defence will be suicide. Comparative embryology seems to confirm recapitulation then it is dangerous.

0. Lack of a recapitulation mechanism. Only repetitions remain.

b. but the understanding of recapitulation is erroneous – too exact. Such a recapitulation has lots of departures. => recapitulation does not exist! => disappointment with the value of ‘Biogenetic Law’ (e.g. De Beer)

3. Haeckel makes problems for us, we don’t like him!Wilkins (2002) to reanimate comparative embryology declares: better that Haeckel never existed but he dedicates a book to Schmalhausen ...

2. Gould sees recapitulation in data, he risks (1977) to save it using heterochroniesbut he loses and ‘recapitulation dies’.

recalls of Darwin:

Schmalhausen:

Changes in early stages of ontogeny cause larger phenotype changes and are typically lethal.

Recapitulations are observed more frequently in more complex ontogeny or organogeny .

It explains ‘Terminal Modifications’ better than de Beer, but not the ‘Recapitulation’.

As programmer I know why and where complex programs grow. This is the lacking mechanism of ‘Terminal Additions’ and in effect – of ‘Recapitulation’.

The first attempt to publish the model– Fifth International Congress of Biomathematics Paris 1975

Mathematical ‘graph theory’ describes networks structure,but the science of networks, which can be complex and can function, has just developed.

nodelink

k=4 node degree

Networks can be:not directed / directed , small / complex, static / functioning ...

Networks, have some propertieswhich are independent of their application

like geometrical figures,

a

h s=ah/2

examples ofnetwork structuresand their elements

linear graph tree feedback

patchinputs

outputs

Networks, have some propertieswhich are independent of their application

like geometrical figures,

a

h s=ah/2

examples ofnetwork structuresand their elements

linear graph tree feedback

patchinputs

outputsOntogeny for de Beer (1940) Severcov (1934) Schmalhauzen (1942).

For me ontogeny (especially conditional specification characteristic of vertebrates) is a complex network of conditional transformations of structuresgrowing randomly under adaptive condition (Darwinian elimination).

The main assumption

functioningcomplexnetwork

environment

Model is designed in 3 steps:

Step 1 - we neglect source of the result.

Step 3 - adaptive condition using small change tendency rejects part of these changes creating structural tendencies(e.g. terminal additions, terminal modifications, growth, recapitulations).

Step 2 - we come back to big picture and changes of result depend on random changes of network (additions and removals of nodes).

Assuming its random changeswe investigate general effects

of adaptive condition- ‘small change tendency’.

requirements

result (properties)

Damage propagationComplexity threshold

Each property has s equally probable variants (1-ideal & rest equally wrong).

Fitness b is a number of identical variants of properties in result and in requirements.

Adaptive condition a ≡ bt+1 ≥ bt (Darwinian mechanism - fitness should not decrease).

For higher fitness b only very small changes of result

can be acceptedby adaptive condition.

Assuming its random changeswe investigate general effects

of adaptive condition- ‘small change tendency’.

structuralmeasure of

D - Depth

output

input

stage of growth

number of nodes

The main result of simulation –Weismann’s ‘Pressing back’

Probable node migration in structure: realtheoretical as was drawn if only +, only in D=0

network

more – than +simplification of deeper parts of network

node changes its depth D

pressing back (down)

early

terminal

late

more + than -terminal addition

=256 nodes

We obtain all 3 elements of Weismann’sconception of ‘Pressing back’ whichsummarizes observations leading toHaeckel’s ‘Biogenetic law’.

functional order

complexity thresholdnetwork at particular time

P(+) = P(-) and is the same in each place of the structure, but not all of these changes are accepted.

complex network

Why balance of accepted + &- is not zeroin different places and totally ?

+ Set of possibilities for + is very large, we must draw at least two nodes (or links) in the network to indicate the place function for node. a. We can always add a new node between network and outputs and increase the fitness (fitness b is never at its maximum value). .b. We can always add a ‘transparent’ node anywhere (b constant) (it does not change network function in its current state).

We reject this possibility in part of simulations using ‘cost’

i.e. a ≡ bt+1 > bt for + and a ≡ bt+1 ≥ bt for -.

- Set of possibilities for – is small, we can only draw a node from the present nodes in the current network.

1. There are different sets of possibilities to draw + & -. + Set of possibilities for + is very large, but - for – is small.

2. Reorganization of conditions for adaptive condition test depends on number of accepted changes in given period and area. Part of tested possibilities in similar conditions + for + is always very small & negligible but- for – is significant and the same possibilities may be tested few times.In network areas with different modification tempo (Terminal modifications P(a|D))P(a|+,D) and P(a|-,D) should differ.

Summarizing:Additions and removals have very different conditions of drawingand testing by adaptive conditions which depend on place in the network.Why does anybody expect identical effects of such different mechanisms?Total growth of a network is also an effect of these differences and adaptive condition.

Parallelism exists or not?

Alternatives (xor) :1. recapitulations – ontogenetic stages are similar to adult ancestors2. repetitions – similarities in ontogeny are its stage relics only

2.Change initiation occurs with equal probability in all length of ontogeny, but it causes changes in all later parts of ontogeny. (de Beer’s mechanism 1940)It explains Terminal Modifications and Conservation of early stages regularity. (Naef 1917)- ontogeny change types include additions and removals similarly,- ontogeny changes occur more frequently near the end of ontogeny.

The choice depends on mechanism of ontogeny changes:1. Haeckel suggests: inheritance of acquired characters

but genetics... It gives Terminal Additions regularity. ( „Pressing back”

conception: Terminal Additions and Compression of

early stages. Weismann 1902), - ontogeny expands by additions of new adaptive transformations at its end.

PhylogenyOntogeny

repetitions

ontogenytime

changeinitiations

Lack of a mechanism. Observations must agree with genetics!

Parallelism exists or not?

Alternatives (or) :1. recapitulations – ontogenetic stages are similar to adult ancestors2. repetitions – similarities in ontogeny are its stage relics (later)

1. ... additions than removals near the end of ontogeny, Terminal Additions

... removals than additions far from the end of ontogeny. Compression of early stages

PhylogenyOntogeny

2.Change initiation occurs with equal probability in all length of ontogeny, but it causes changes in all later parts of ontogeny. (de Beer’s mechanism 1940)It explains Terminal Modifications and Conservation of early stages regularity. (Naef 1917)- ontogeny change types include additions and removals similarly,- ontogeny changes occur more frequently near the end of ontogeny. repetitions

ontogenytime

changeinitiations

Common assumption (as recapitulation opposition)Change occurs with equal probability in all length of complex ontogeny,change types include additions and removals similarly, but Darwinian elimination accepts more ...

2. ... changes near the end of ontogeny, Terminal Modifications than far from the end of ontogeny. Conservation of early stages Darwin’s mechanism Naef (1917)

Gecow’s mechanism Weismann (1902)

Mechanism of Terminal Additions needs Terminal Modifications.

It creates Recapitulation and Biogenetic Law which exist in theory.

- Yes. We should expect it in nature.

Modification and Conservation of Deeper Parts of Network

Tendencies of Terminal Predominance of Additions and Simplification of Deeper Parts

Tendency of Terminal

deep fade out

on k=1,0

Similarity of functional D and historical H orders

Similarity of functional D and historical H orders(connections to the network)

Thanks for your attention.Viva Recapitulation and Biogenetic Law!

Down with disbelief and blindness!etc.

Ontogeny is a complex network of conditional transformations randomly evolving under adaptive condition. This is sufficient to expect Weismann’s Terminal Additions and Haeckel’s Recapitulation.

Viva Haeckel and Weismann!Viva Comparative Embryology!

[email protected]/people/Gecow/Gecow.htm

Gecow A. 2005. From a “Fossil” Problem of Recapitulation Existence to Computer Simulation and Answer. Neural Network World 3/2005, 189-201

Gecow A. 2007 (?) Structural Tendencies - Effects of Adaptive Evolution of Complex (Chaotic) Systems. Int.J.Mod.Phys.C, in press.

Gecow A. 2007 (?) Emergence of Growth, Complexity Threshold and Structural Tendencies During Adaptive Evolution of System. EPNACS conf. in ECCS’07 Dresden