The Theory of Complex Accumulations

1. In Evolution, some attributes are hard-to-evolve not because they result in complex organs, but because they accumulate many low-yield fitness gains over long periods. This requires unique conditions for the new attributes to evolve.
2. Yet once a new, complex attribute (such as vertebrae, feathers, or a four-chamber heart) has evolved this way, it can close off conditions under which a similar attribute might evolve again. (In most conditions individuals cannot afford changes of genotype that might only yield very low gains.)
3. Once a complex accumulation stabilizes the host species tends to radiate into many varieties. New varieties in turn make it harder for a similar novelty to evolve in rival types against the new, more highly evolved types.
4. More controversially, as complex changes accumulate, host types will loose absolute fitness, despite that each change was a small positive fitness gain.

(Note: This page explains why fitness falls, both absolutely and relatively. Please also do not leave without reading 2.3 Complex Accumulations. If you disagree with any assertion, please tell me why. mailto:[email protected] )

(Back to Top.) ____________________ (Site Navigation Page.)

2.1 Fitness and Complexity

2.2 Absolute Fitness

2.3 Complex Accumulations

2.4 Stepped Changes in Evolution

Return to A New Model of Evolution Home Page

No principle of evolution appears more self-evident than that individuals will enhance fitness by accumulating complexity. This is how evolution seems to work. An individual accumulates a slight improvement. This might make it a more complex creature, but if it enables it to procure more offspring than its rivals fitness goes up. (Individuals do not accumulate genetic complexity in single lifetimes. If among a group of individuals those born with more complex attributes are reproductively favored in the struggle for life, over time life will evolve more complex forms.)

This idea matches human expectations.

1. Humans often see themselves at a pinnacle of a scale of complex beings.
2. So when Darwin showed that evolution was a "struggle", it fitted that creatures were "struggling" to become more complex (more like us).
3. While there are scientific and emotional flaws to these views, we still accept the syllogism that the "fittest" struggle to "evolve", by becoming more complex.

Moreover, if we define the increase of complexity as accumulation of genetic distance in the direction in which time flows, and treat fitness as a genealogy over many generations, fitness will increase in the direction of accumulation (the direction of time) virtually by definition. (We think of the history of life as the history of organisms becoming progressively fitter.)

Yet increasing complexity might also carry a fitness cost, although this depends on the definition.

• Usually, we mean by fitness that an individual is well adapted. But this is subjective without a means to quantify it.
• Fitness should pass some objective test, such as ability of an individual to survive, attract mates and procure offspring. If an individual does this, its unique DNA should end up widely distributed among descendents, which would be proof that the individual was fit.
• We test this by measuring the number of unique genes (or unique DNA, it is not the same) that an individual will pass into subsequent generations.

This test makes fitness a retrogressive rather than an immediate measure, but that is the price of objectivity. We cannot observe a strong or fleet individual and assume that attributes aesthetically pleasing to humans are fit biologically. We must measure the results to be sure. We can only measure results among subsequent, rather than immediate generations.

If we can assess fitness objectively by measuring the amount of ancestor DNA distributed among descendents, the greater the number of generations we measured this over the better the result should be.

• If we were examining a single trait, present in an ancestor but retained unaltered thousands of generations later, the trait was fit.
• If we examined an individual genome, including its non-genetic DNA over a few generations, and the individual DNA fingerprint is more widely distributed than its rivals, that individual was fit.

Yet, while viable traits can persist unaltered over thousands of generations, unique genomes, including the non-genetic and non-expressed DNA become quickly diffused as types change.

1. If an ancestor accumulated a new novelty that enhanced the ancestor’s ability to procure offspring, this increases fitness, by allowing the ancestor to propagate its unique DNA into subsequent generations.
2. But it would only increase fitness if the first generation acquired a new novelty. If individuals of later generations also acquire new novelties, ancestor fitness would eventually fall!
3. Each generation would be phenotype fit (the adult had offspring), but after many generations of small changes the ancestor's unique DNA would slowly diffuse. In each generation small changes would "vector" the modern genotype a large genetic distance from the ancestral one. (There would be less of the ancestor's genes in the modern genome after many generations.)

For example, suppose an ancestral fish acquired a small novelty that made it a better fish. Then that ancestor was very fit, because if the design is sound fish species can survive unchanged for huge periods. But suppose that for less well-adapted fish the ancestral seas were already saturated with fit designs. So, the only hope of procuring a lineage lay not in the sea, but pushing onto land (by becoming an amphibian). Amphibians did penetrate land so this was a fit phenotype move. But that did not make the ancestor genotype fit over many generations.

• If an amphibian is competing to be a better amphibian, and it acquires "better amphibian" genes, its fitness will rise in competition with other amphibians, but only over a short distance.
• Over a long distance, the amphibian began as a fish and fitness measures how much unique DNA an individual propagates through subsequent generations.
• The fish ancestor lost fitness when its lineage diverged from being fish and became amphibian, despite that becoming amphibian was fitter than going extinct.
• Similarly, the amphibian ancestor would lose fitness when its own lineage diverted further from "amphibian" to become reptile, although this is a fit lineage move, because a reptile is a more stable design than an early amphibian.

So, for each small step on the way from fish to reptile, for each fitness plateau of a stable intermediate species, fitness will rise. But over the entire ‘distance’ from fish to reptile, fitness will fall by a huge amount, which is what the new theory is saying. Fig 2.2 shows the problem. G3-G4, although only a transitory type, was stable enough that "G3" was a fit ancestor (to generations G4 and G5). But once major "vectoring" began again at G5, ancestor "G3" was no longer fit as an ancestral genotype.

Still, this is showing fitness falling as a relative measure. But does fitness also fall in absolute terms? In some ways, this question has no meaning, as fitness, ultimately, is a human measure of how we understand nature. But there is evidence that fitness does fall in absolute terms as complexity increases.

Fitness, always, is a human measure. In nature organisms do what they do, but modern humans, the intelligent creatures, care that what organisms do makes logical sense. One powerful method for humans to understand what organisms do is measure the amount of unique DNA than an organism, by its behavior, will pass into subsequent generations. Only this is not an easy measurement to take.

• No technology allows us to measure every DNA combination an organism will pass on.
• For extinct species, we can only measure DNA fitness inferentially, among modern descendents of lines that survived.

So most fitness modeling is instead done mathematically, and results compared to that observable among living species. These models show that if a modification to genotype improves adaptability, fitness will rise. Only such models can only be applied conclusively over a few generations, within the same species. But evolution has proceeded over billions of generations and millions of species. Over such huge distances, for what humans perceive as "improvements" (increase in complexity) fitness will fall. We can demonstrate (Fig 2.2) that fitness will fall as a relative measure, against an ancestor from a line that widely diverged from original type. But there is evidence that fitness falls absolutely for increase of complexity too.

• Sexual reproduction is more complex than earlier evolving asexually reproduction. But for reproduction by sex genotype fitness falls by 50%, over 100% for the less complex asexual forms. (Fitness falls by 50% of DNA unique to an individual. But expressed genes can be less than 3% of total DNA, and mostly these will be 100% conserved. See Smart Gene Hypothesis.)
• We suspect (there are no agreed ways to measure it) that earlier, prokaryote life, is reproductively fitter than more complex eukaryotes, which evolved later. All prokaryote sequences are selected, which restricts the genome "personality" to fit sequences only. But eukaryotes include many (up to 97%) non-selected sequences. This gives each eukaryote a more distinctive genome, but it is also one more quickly altered or diluted (less exact copy of total genome) in subsequent generations.
• Eukaryotes contain symbiotic genomes of other prokaryotes (like mitochondria genes). This forces the original eukaryote genes to "share" reproductive success with immigrants, plus it gives the eukaryote cell less control over the total DNA transmitted at each generation. (Many mitochondria genes are also non-selected.)
• Prokaryotes reproduce more often and in greater total numbers, and it easier to exactly copy a genetically simple organism, than a complex one.
• If fitness means retaining stable designs over very long periods, simpler, earlier evolving organisms like worms, sponges, or mollusks, or cyanobacteria seem to do this best.

There is also a theoretical argument that fitness must fall from early life, due to so-called "minimum complexity". There must be a minimum level of complexity for any DNA-coded cell to exist. Yet if there is a minimum state of complexity living reproduction can exist at, then any descendents at that state must be exact copies. There can be non-exact copies, but these would only add complexity to the minimal possible amount. We doubt if any primitive cell ever existed at minimum theoretical complexity. (A perfect state of anything is too unlikely in a heterogeneous universe!) Yet, the fittest organism ever was probably a cyanobacteria, which hardly changed its genotype in 3 billion years. But if this organism existed first, and had high fitness, and organisms that evolved later enjoyed less fitness, at some point organisms must have evolved against a loss of fitness.

Moreover, if fitness is passing of unique parent DNA into subsequent generations there are many reasons why fitness would fall as complexity increased. To suggest a few;

1. Genomes of complex organisms do not compose a high set of "new" genes. Humans say, share 98% of their genes with chimps, and over 99% of their genes with each other. So, whereas new genes in ancestral algae could highly propagate throughout subsequent life, such high fitness is not available to complex organisms. Even if a human were born with such a fantastic genetic advantage that his lineage would eventually replace all existing human lines it would still only influence less than 1% of the total genome of life on earth, for countless generations.
2. Genotypes are a reproductive "client" to each gene composing them. But complex genotypes do not service the reproductive needs all their genes equally well. The 98% of genes that human and chimps share did well, but the 2% of genes which can be varied for dramatic changes of behavior and morphology have low fitness, and can quickly be changed or deleted from the gene pool. So, among organisms with high variability and rapid rates of evolution genes for homologous sequences do well, genes (or alleles of polygenic traits) for variable sequences can suffer very low rates of fitness.
3. Fitness rewards for simple organisms can be very high, such as trillions of generations with little alteration. Such rewards do not exist for complex organisms. Larger size, longer generation periods, much fewer offspring, fiercer competition, limited ranges, greater environmental exposure (to asteroid strikes, etc.) all limit the potential fitness of late evolving, complex organisms. (Complex organisms, such as humans, can enjoy high ‘type’ fitness, just as mammals can enjoy higher type fitness today than reptiles. But the issue here is fitness of a single individual. A unique reptile genotype has greater chance of propagating with less total alteration than a mammal’s genotype, though there will be exceptions either way.)
4. Complex organisms consume more energy to live and reproduce, and must survive at the end of more complex and delicate food chains. Life and reproduction are "harder" as basic physics as life grows more complex, though complexity yields other options. (One could write an entire book exploring this argument, so we will not discuss it further here.)

Only if fitness does fall as organisms become more complex, then we must find the reason. It is because in evolution there are both hard and easy to enact changes.

• Easy-to-enact changes that enhance adaptation will always result in a fitness gain.
• Hard-to-enact changes can only evolve against a fitness loss.

Explaining why is one of the most difficult to grasp of all modern concepts of evolution.

An early concept of evolution, partially adopted by Darwin, was that evolution was progress towards perfection, usually understood to be evolution of intelligent beings. Yet, modern studies show evolution is mostly non-progressive. Changes are more like oscillations than progress towards a goal. While ever organisms maintain or slightly improve fitness, this is the case. Again, it is because changes are always of two types.

1. Analogous, or as we call them here, polyphyletic changes, such as varying length, shape, texture, color, or behavior, are easy-to-enact and occur all the time.
2. Homologous or monophyletic changes, for new materials, technologies or body plans (tissue, hearts, nerves, lungs, vertebrae, limbs, etc.) take a huge evolutionary effort and millions of generations to perfect.

Now why, in primal life, some bio-chemical properties are stable but hard to change and some properties are less stable but easy to vary, concerns the physics of the universe. But once life advances to DNA-coded multi-celled organisms the reason that some changes are difficult is statistical. Some changes will require large alterations to genome, many of which will be statistically unlikely to yield high fitness gains in a single generation. Only this argument is controversial. Opponents of evolution have argued that the eye, say, could not evolve because too many changes are required at once to produce the perfected organ. But while it contains hard-to-evolve attributes (muscles, nerves, photoreceptors) the eye assemblage is not itself a complex accumulation. It has been shown (many times) that each incremental improvement in vision, even for a partially perfected eye, always yields a slight fitness advantage. Plus eyes are not monophyletic. Several types of eye have evolved several different times, so even if we had no clue how eyes evolved, it must still be by a simple fitness path, because eyes evolved so many times.

But take a genuinely monophyletic attribute, such as the four-chambered heart.

1. Hearts in fish have two-chambers, with an undivided auricle and ventricle.
2. In amphibians the auricle becomes partially divided, and in reptiles this extends to a partial division of the ventricle.
3. Mammals have full division of each auricle and ventricle into a four-chamber heart.

Although it sounds involved, growing the tissue dividing the chambers is not itself complex, as even a part divided chamber will give a more efficient blood flow than no divide at all. Yet despite that it appears a simple adaptation to fully divide the ventricle, no creature other than a mammal has ever done so, or ever will. The reason is that while a fully divided ventricle will yield more efficient blood flow, it will not yield an overall fitness advantage without extensive modification to lungs, skin, arteries, breathing, and endothermic regulation. This means that he four-chamber heart comes not as an isolated improvement, but part of a repertoire of modifications to a large number of attributes, each of which must work with the other to yield a high fitness gain. Only each attribute has its own slight improvement, and in competition with a neighbor, even the slightest improvements are passed on. So, although we can see why these attributes would also evolve in isolation, we must explain why they evolved only once.

Again, it has to do with fitness, and that some changes are easy, but others are hard to evolve. Take the case of reptilian predators, also in competition with each other, for a very slight advantage. Now extending the length of anything is an easy-to-enact change. In a single generation, one reptile might produce a slightly longer claw, the other a slightly extended ventricle divide. Both improvements can be produced, but under the prevailing conditions of competition, which is more likely to be fit?

• There are conditions where a slight extension of the ventricle will be a fitness improvement, but these are more likely where other improvements (breathing, skin, endothermy, etc.) are already under way.
• But most individuals, in most species, do not have the opportunities to experiment with complex improvements. Evolution is "the quick and the dead". Improvements such as a slightly longer claw, sharper tooth, streamlined shape, or improved vision can yield quick, single generation fitness advantages. Most individuals, in most species will be confined by fitness to these quick, "fitness-rich" enhancements.

The four-chamber heart is a complex accumulation them, but not because it is complex to extend the ventricle divide. Rather, extending this alone yields low fitness in isolation from other changes, compared to self-supporting "fitness rich" adaptations an individual can undergo. Yet all attributes that are monophyletic (evolved at one source) and homologous (radiated from the source into many types) are complex accumulations. Each attribute has a small fitness gain. But the big gains are long term, and involve modifying many features, which requires evolution over a huge genetic distance. But as we also saw, large changes of genome eventually result in fitness loss, as a relative measure for ancestors, but fitness also falls absolutely for modern types, for many reasons. To these reasons we can now add two more.

1. Low-fitness yield enhancements are only favored once individuals are "pushed" down a slope of general fitness decline. (It will be explained. See Peripheral Niche Mechanism next section.)
2. Complex accumulations once they mature, invariably result in large radiations into many new types. So as new types emerge ancestors suffer further relative fitness loss, while modern types have to compete among greater diversity in greater complexity and suffer absolute fitness declines.

Note: The concern over fitness loss among ancestors is not over the "feelings" of the ancestor, who has long perished. A "retrogressive" fitness loss means that a large portion of the ancestor's genes turned out to be not suitable for modern conditions. This forces the evolutionary cost of further adaptation onto descendents.

Contention over whether evolution proceeds by continuous small steps or consists of long periods of stasis "punctuated" by sudden increases in species complexity is now well debated. However, most issues concern interpretation of the fossil record and relative rates of evolution. Yet, no matter which view one holds, sudden, large mutations at many alleles at once are statistically unlikely to be fit in a single generation. So, if evolution does proceed in a stepped manner, then small changes accumulating over long periods must somehow be causing the stepped effect.

We can show stasis and the "stepped" changes as an effect of fitness, again to do with all changes in evolution being either easy or hard to evolve.

1. If easy to change adaptations (length, size, sharpness, etc.) are available individuals will be forced to consume these first.
2. Fiercely competing individuals struggling for the slightest advantage cannot afford experimentation with complex new novelties that require large changes of genome and long evolutionary times to perfect. (This is "retention". It means that in stable competition, no individual can afford to deviate too far from type without loosing fitness.)

For example, in hundreds of millions of years no descendent of a shark could afford the fitness penalty of evolving into anything other than a more efficient shark. The ancestral shark design was already very fit, and sea conditions are mostly stable, so fitness among quick, easy-to-enact, single generation improvements, retains a shark always to be a shark.

Yet while it never happened to sharks, many events can temporarily "loosen" the strict fitness bind that sharks say, are always in. In stable conditions, not only a single species (sharks) but all the species within an environmental range will be in fierce competition for very slight advantage. This creates a range "equilibrium" in which each individual in each species in the range can only afford easy-to-enact fitness enhancements to stay ahead. (It is not certain if this is what Steven Gould meant by equilibrium. It is never this sharp, but easy and hard to enact changes form an on-off, either-or, "state". So species within a range being all restricted to easy-to-enact changes is a "state" of equilibrium, defined perhaps with a sharpness of distinction no biologist could accept.)

However, the equilibrium of any range can be "punctuated" by unsettling events, like environmental or geological change, vacation of ranges by extinction of rival species, invasion of ranges, overpopulation, migration to new ranges, and so on. When this occurs the fitness rules become loosened.

• In an expanding range fitness penalties will not be as severe. (There will be plenty of resources.)
• In a contracting range prior penalties are no longer working anyway, so fresh novelties and new means of survival must be found.

Loosening of fitness rule favors complex accumulations, because these consist of many small, low fitness yield changes. But although many changes are required they will tend to be in spurts. While genotypes are changing and fitness yields are low (yields are low because types are changing) the species will be "searching" for a new equilibrium plateau, where high gain, easy-to-enact yields will again predominate. (Roughly, the founder individual who adapts to the new plateau best will enjoy high fitness. But until the new plateau is found, no ancestor will enjoy high fitness because every few generations the founder type must alter again, until the process stabilizes.)

Stepped changes in evolution are not hard to explain, once we accept that all changes are either easy or hard to enact changes, and that fitness falls as species become more complex.

• The stasis or equilibrium "state" is when all existing types within a range are so optimally adapted that individuals can only afford easy-to-enact, fitness rich changes, to stay ahead.
• Loss of equilibrium occurs when this state is upset by external factors, and (individuals in) species are forced to "search" among previously low fitness yield advantages, for fresh enhancements that will adapt them to a new, high yield fitness plateau.

But if fitness yields for changes are low, and conditions have not stabilized, there will be much experimentation and many changes until a new stability is found. This will appear in the fossil record as a "stepped" or relatively rapid rate of change, although there are other ways to explain this. (Again, this is phylogenic explanation of the stepped effect. There can be random changes of climate or geology, or migration and extinction for non-phylogenic reasons.)

Contact Information