Evolvable soma theory of ageing: observations from the natural world

Abstract

Ageing has traditionally been interpreted through declining selection gradients with age. The Evolvable Soma Theory of Ageing (ESTA) provides a complementary perspective by proposing that age-associated somatic changes reflect the late extension of the developmental program, which is shaped by evolution. These late-acting changes are epigenetic in nature, encoded in the genome, subject to germline-mediated modification, and may be viewed as developmentally regulated variants that have not yet been refined by selection. Here, we explore whether patterns observed in the natural world are consistent with key expectations derived from this framework. One such expectation is that species experiencing lower rates of evolutionary change may exhibit slower age-associated decline. In line with this idea, we examine long-lived species occupying ecological niches characterised by reduced competition and environmental turnover, as well as taxa often described as exhibiting high evolutionary stasis, and note that several display extended lifespans compared to related groups. Secondly, we examine evo–devo dynamics concerning the relationship between development and evolutionary history, and discuss their compatibility with the conceptual structure of ESTA. A third expectation is that stronger evolutionary pressure may influence life-history timing, including shifts in the age of sexual maturation; this possibility is explored through comparisons between bonobos and chimpanzees and through broader sex-specific patterns reported in mammals. Taken together, these observations do not constitute definitive tests but provide preliminary, qualitative support for the idea that evolutionary dynamics and ageing may be more closely intertwined than traditionally assumed, and they highlight directions for future investigation.

Introduction

Despite decades of research, the mechanisms of ageing remain elusive, with the relative roles of genetic and non-genetic factors still debated. Stochastic theories [26] attribute ageing to the gradual accumulation of damage from environmental stressors or metabolic by-products, much like the wear and tear of mechanical systems. In contrast, programmed theories [3] view ageing as the outcome of a genetically regulated process, though its progression may be modulated by external factors. Epigenetic theories [23] emphasise changes in gene expression patterns, rather than alterations in the DNA sequence itself, as the principal drivers of age-related dysfunction and tissue decline.

If ageing is a mystery in its own right, its relationship with evolution is even more perplexing. The prevailing theory posits that ageing persists because genes that are activated later in life experience diminished evolutionary pressure, a concept known as the Mutation Accumulation theory [27]. From this viewpoint, ageing may result from trade-offs associated with genes that have pleiotropic effects, as outlined in the Antagonistic Pleiotropy theory [47]. Alternatively (or additionally), ageing can be understood as a byproduct of limited investment in cellular repair and maintenance, prioritising early-life evolutionary advantages, a notion encapsulated in the Disposable Soma theory [25].

Another perspective frames ageing as an evolved trait (an “evolution of ageing”) that may confer benefits to the individual or the species. August Weismann originally suggested [44] that natural death evolved to clear space for younger generations, a view he later abandoned and that is now considered incorrect. Kin selection theory [45] posits that post-reproductive behaviours can enhance inclusive fitness by improving the survival prospects of related offspring. In line with these ideas, [18] argued that limiting lifespan may prevent older individuals from monopolising reproduction, thus maintaining a population’s evolutionary adaptability. A recent model [36] proposes that evolution has shaped mechanisms linking fertility timing and ageing, making ageing populations more evolvable and successful.

This work is concerned with a recently proposed theory [13], which conceptualises development and ageing as a continuous process, driven by epigenetic changes in target cells, controlled by genetic instructions encoded in the zygote’s genome and subsequently inherited by all cells. In this context, ageing reflects the cumulative manifestation of epigenetic changes expressed during the post-reproductive phase, which are not yet evolutionarily optimised but are instead subject to ongoing selection, functioning as somatic “experiments” through which evolution explores novel phenotypic variation. These experiments are often detrimental, leading to progressive functional decline and eventual death, while a small subset may produce beneficial adaptations that evolution can exploit to shape future developmental trajectories. In this work, we gather evidence in favour of this theory through observations from the natural world.

The ET model

The ageing theory examined here builds on the Epigenetic Tracking (ET) model, in which a distributed population of stem cells orchestrate development by triggering change events that affect normal cells. The cell genome contains two classes of information: basic genes, which encode structural building-blocks, and developmental genes, which specify when and where change events occur. These events modify the epigenetic state of surrounding cells, leading to downstream outcomes such as differentiation, proliferation, or apoptosis (Fig. 1). The temporal component of events is controlled by gene timers associated to developmental genes, acting in synchrony with a global clock shared by all cells. Full model details are provided in [13].

Fig. 1.

The genome is organised into basic and developmental genes. Developmental genes encode change events that progressively remodel the epigenetic landscape of specific target cells. In the example shown, three developmental genes are activated in stem cells B, C, and A at clock values 2, 3, and 4. Each gene specifies both the target cell population and the epigenetic modifications to be applied to their initial state. These modifications result in the structural enabling or disabling of certain basic genes (bottom right of each panel), leading to altered cell behaviour, proliferation, or apoptosis. Together, the sequence of change events shapes development

Basic genes may correspond largely to protein-coding genes, whereas developmental genes could involve long non-coding RNAs (lncRNAs), now recognised as key regulators of epigenetic information [31, 35]. Epigenetic modifications encompass a range of chemical changes that coordinate gene activation in time and space, including writing, erasing, and reading elements of the “histone code” [24] and the deposition or removal of DNA methylation marks. This regulation drives the emergence of specialised cell types during development, as epigenetic mechanisms guide cells through successive steps of differentiation, and the genetic pathways that orchestrate these epigenetic dynamics are now beginning to be elucidated [49].

In the ET model, development proceeds through a series of developmental steps, each associated with a specific clock value and hosting a number of change events. When combined with a genetic algorithm that simulates evolution, this framework reproduces an evo-devo dynamic capable of generating complex structures [12]. A simplified form of kin selection is incorporated [45]: individuals are favoured if they maintain fitness for some time after reproduction, reflecting the biological importance of parental care or investment in kin. Operationally, an individual passes through N developmental steps, with fitness evaluated across multiple steps after the designated reproduction step r, weighted so that early post-reproductive steps contribute more to fitness than later ones.

Developmental genes scheduled to activate at steps $>r$ (during the ageing period) contribute progressively less to fitness and are therefore subject to weaker evolutionary optimisation. As a result, the events they trigger exhibit a form of pseudorandomness: although encoded in the genome and thus deterministic, their effects appear increasingly random. This pseudorandom character increases the probability that such events are detrimental rather than beneficial. Therefore, within the ET framework, development and ageing are understood as phases of a single continuous process: development is driven by optimised developmental genes, whereas ageing arises from non-optimised developmental genes that become active in specific stem cells after reproduction [16].

The ESTA hypothesis

Understanding the biological role of ageing requires recognising that evolutionary changes arise from random variation, and since random alterations are statistically more likely to be harmful than beneficial, one would expect the effect of detrimental changes to be evident in all evolving organisms. The Evolvable Soma Theory of Ageing (ESTA) proposes that ageing represents precisely this phenomenon: the cumulative effect of genetically encoded modifications that become active in the post-reproductive phase. These late-acting modifications are not yet optimised by evolution but are instead subject to ongoing optimisation, serving as somatic “experiments” through which evolution explores novel phenotypic variation. Most such experiments are maladaptive, producing functional decline and ultimately death, whereas a minority may yield advantageous traits that can be co-opted into future developmental pathways.

ESTA foresees the incorporation of novel traits in later-life stages, with minimal impact on overall fitness and reduced risk. Successful innovations can then be shifted to earlier developmental steps through adjustments in gene timers, thereby stabilising their integration. For example, evolution may generate late-acting mutations that influence cardiovascular function; most will be harmful, contributing to hypertension or heart failure, but occasionally a mutation may protect the heart and vessels in old age. Individuals carrying such a beneficial variant would remain healthier for longer and could increase their reproductive success directly or indirectly through kin support. Their descendants, inheriting the same variant, would gain a selective advantage, making the change more likely to spread. Over generations, such advantageous late-life innovations could move earlier in development, enhancing fitness more substantially.

For evolutionary forces to act on the developmental program, its components must be modified in the germline or at the zygotic stage. Mutations can occur at any parental age and for many reasons (for example, through exposure to ionising radiation [42]). If such mutations affect developmental genes, they may alter the change events these genes control. One of ESTA’s assumptions is that the mutation frequency is linked to developmental genes’ timer values, with late-acting developmental genes exposed to higher mutation frequency than early-acting ones [14, 15]). By contrast, mutations arising solely in somatic cells are not transmitted to offspring and thus remain invisible to evolution, although they can still contribute to ageing through their cumulative effects.

ESTA can be contrasted with the Mutation Accumulation, Antagonistic Pleiotropy, and Disposable Soma theories, here collectively referred to as Standard Evolutionary Ageing Theory (SEAT). In brief, SEAT views ageing as a byproduct of waning evolutionary engagement, whereas ESTA interprets ageing as evolution in progress. Both perspectives agree that the fitness impact of genes decreases with later expression, but they differ in conclusion: SEAT argues that evolution acts mainly on early stages to maximise impact, while ESTA posits that it acts on later stages to minimise risk. Put differently, SEAT focuses on past evolution, treating ageing as neglect, while ESTA highlights present evolution, framing ageing as the outcome of ongoing experiments required for innovation.

The ESTA model shares some conceptual elements with the “quasi-programmatic” framework proposed in [7], but there are important distinctions. In Blagosklonny’s view, quasi-programs represent developmental processes that persist into old age and are expressed “per inertia”. By contrast, ESTA envisions a unified developmental program that extends continuously from conception to death, composed of instructions with specific spatio-temporal activation patterns. In this context, late-acting events are not residual instructions carried over into the post-reproductive period, but distinct instructions that remain dormant during development and are programmed to activate later in life.

Another key difference between Blagosklonny’s model and ESTA is that in the former ageing is propelled by genetic mutations (in protein-coding genes), while in the latter, it is driven by epigenetic changes under genetic control (possibly encoded by lncRNA genes). Finally, Blagosklonny’s framework does not attribute any evolutionary significance to late-life quasi-programmatic hyperactivations and remains, therefore, consistent with SEAT. ESTA, on the contrary, sees ageing as the effect of evolution in progress: while most “experiments” fail, some may introduce phenotypic novelty that selection can refine across generations.

Some ESTA’s predictions

In previous work [14, 15], the connection between evolution and ageing proposed by ESTA was explored through computer simulations based on the ET model. These simulations suggested that ESTA achieves greater evo-devo efficiency than SEAT, with faster fitness progress during evolution and a more gradual post-reproductive functional decline in development. The analysis also indicated that the core features of ESTA (distributed fitness evaluation, timer-dependent mutation rates, and evolvable gene timers) are central to its enhanced performance. It should be emphasised, however, that the ET model represents a substantial simplification of biology: more sophisticated and biologically detailed simulations will be required to validate these results.

In the present work, we seek evidence from the natural world, focusing on three key predictions of ESTA. The first prediction follows directly from the idea that ageing reflects evolution in progress: species with faster evolutionary rates should exhibit faster ageing, whereas species with slower evolutionary rates should age more slowly. Where evolutionary innovation is strongly required, the genome is expected to accumulate a larger number of non-optimised late-acting developmental genes, increasing the frequency of “failed experiments” and leading to a steeper functional decline after reproduction. Conversely, in contexts characterised by reduced evolutionary pressure, fewer late-acting genes persist, failed experiments become rarer, and functional decline proceeds more gradually. This relationship should be particularly visible in species with exceptionally low evolutionary rates.

The second prediction relates to the classical recapitulation hypothesis, often loosely associated with Ernst Haeckel’s phrase “ontogeny recapitulates phylogeny.” Historically, this idea suggested that embryonic development passes through stages that in some ways resemble the adult forms of a species’ evolutionary ancestors. While Haeckel’s original formulation is no longer accepted, the general notion that developmental stages may reflect aspects of evolutionary history retains heuristic value. From the perspective of ESTA, a more nuanced interpretation is proposed: evolutionary innovations may initially appear during later developmental stages, where they can be “tested” in the context of ageing, and in some cases, elements of these innovations may subsequently shift to earlier stages in later generations. This prediction is evaluated using insights from previous computer simulations.

The third prediction concerns a long-term effect of evolution. When evolutionary pressure is high, evolution proceeds more rapidly, and ageing accelerates. Over generations, new developmental stages accumulate at the end of development, consistent with the principle of terminal addition. This process shifts the age of sexual maturation (which marks the end of development) progressively later, while steepening the post-reproductive functional decline. The effect can be quantified by the parameter $LM=(L-M)/L$, where L is lifespan, and M is age of sexual maturation; stronger evolutionary pressure is expected to produce lower LM values, whereas weaker pressure should yield higher values. This prediction is assessed using an example of closely related species, as well as patterns emerging from sex-specific differences in mammals.

Longevity and environmental competition

The ESTA model predicts that the ageing rate of a species is proportional to its evolutionary rate. In this section, we assume that the evolutionary rate is directly proportional to the selective pressure exerted by environmental competition. This competition encompasses inter-species dynamics (e.g. avoiding predators, countering prey defences), intra-species rivalry (e.g. securing resources and mating opportunities), and environmental changes. Higher levels of competition in any of these domains increase the demand for evolutionary adaptation, thereby increasing the need for evolutionary experimentation and leading to faster ageing. On the contrary, lower levels of competition should yield lower ageing rates.

Since most species experience substantial evolutionary pressure and measurable ageing, the predicted correlation is often obscured. The strongest signals are expected in exceptional environments where competition is minimal and evolutionary pressure is greatly reduced: in such cases, longevity should become more pronounced. According to ESTA, the reduction in evolutionary and ageing rates is not (primarily) achieved through a small set of genetic variants. Rather, it stems from a gradual loss in the genome of developmental genes that are active during the post-reproductive phase. This is the part of the developmental program under evolution and responsible for ageing: when evolution is less needed for survival, this part will progressively, over the course of generations, disappear, along with its associated ageing effect (scenario B in Fig. 2).

Fig. 2.

Life can be viewed as unfolding in two phases: development, driven by developmental genes already optimised by past evolution, and ageing, driven by genes still undergoing optimisation by present evolution. The timers associated with developmental genes are reflected in their position along the developmental time axis, and their degree of optimisation is represented by the green shading (darker indicates more optimised). Scenario A (baseline competition): many late-acting genes remain active, producing failed experiments that reduce performance. Scenario B (low competition): fewer late-acting genes persist, so failed experiments diminish, together with their ageing effects, and performance experiences a milder decline

The conditions required for this scenario are frequently met in ecological niches. Many exceptionally long-lived species occupy highly specialised, stable environments with reduced ecological competition and abundant resources, factors that consistently support extended lifespans [30]. For example, the bowhead whale (Balaena mysticetus) can live for over 200 years [46]. This species faces no natural predators due to the absence of larger fish in its habitat. Finding food is effortless, as it can simply consume plankton by filtering it through its open mouth. Interspecies competition is likely minimal, and the environment, defined by the cold depths of the Arctic and sub-Arctic oceans, has remained stable and largely unchanged for hundreds of thousands of years.

Another long-lived species is the naked mole-rat (Heterocephalus glaber), whose lifespan can exceed 30 years, remarkably long for a rodent of their size [8]. The naked mole-rat occupies a unique ecological niche in the underground burrows of East African deserts, where it has adapted to low-oxygen environments and a diet primarily consisting of tubers [37]. These subterranean rodents exhibit eusocial behaviour, living in colonies with a single breeding queen and non-reproductive workers, a social structure rare among mammals. Their burrows provide a stable environment with few predators, contributing to their extended lifespans.

The Tuatara (Sphenodon punctatus) is a reptile endemic to New Zealand [9]. Although it closely resembles lizards, the Tuatara is the sole surviving member of the Rhynchocephalia, an order once highly diverse. This fascinating creature occupies a stable ecological niche within its native habitat, benefiting from an abundance of prey such as insects and small invertebrates. One of the Tuatara’s remarkable adaptations is its ability to significantly slow its metabolism during periods of food scarcity, allowing it to survive for extended periods without food, which is particularly advantageous in the fluctuating environments of New Zealand. The Tuatara is known for its impressive longevity, with individuals capable of living up to 90 years or more, making it one of the longest-lived reptiles in the world.

The Greenland shark (Somniosus microcephalus) is one of the longest-living vertebrates on Earth, with some individuals estimated to be over 400 years old. Native to the deep, cold waters of the North Atlantic and Arctic Oceans, it can grow up to 7 m in length and has a sluggish, slow-moving nature adapted to its frigid environment. Despite its lethargy, it is an apex predator, feeding on a variety of fish and marine mammals, and is even known to scavenge carcasses, including those of polar bears and whales. Due to its large size and deep-sea habitat, the Greenland shark has few natural predators, although humans and orcas may pose occasional threats. Its long lifespan was determined through radiocarbon dating of eye lens proteins, as shown in a landmark study [29].

The Galapagos tortoises (Chelonoidis nigra) [41] inhabit the unique ecological niches of the Galapagos Islands, where they benefit from a relatively stable environment with reduced competition for resources. The islands’ isolation has allowed the tortoises to evolve distinct adaptations, such as their large size and herbivorous diet, primarily consisting of grasses, leaves, and cacti. This reduced competition, combined with their slow metabolism, enables Galapagos tortoises to thrive in their habitats, playing a crucial role in the ecosystem as grazers that help maintain the balance of plant life. The Galapagos tortoise is a remarkable species known for its impressive longevity, often living over 100 years and sometimes reaching ages of 150 years or more.

Longevity of living fossils

As outlined in the “Longevity and environmental competition” section, one of ESTA’s key predictions is that higher longevity should be associated with greater evolutionary stasis. To test this, we examine species often described as living fossils, a term introduced by Darwin to characterise organisms that have changed little over long evolutionary timescales. Quantifying evolutionary stasis remains difficult, however, since the term lacks a rigorous and universally accepted definition. Despite these limitations, the notion of living fossils can still serve as a useful, if imperfect, proxy for stasis, and provide a practical means of identifying species that illustrate the predicted correlation between longevity and evolutionary stasis.

In this study, we employ the framework developed by [6], which introduced the Evolutionary Performance Index (EPI) as a scalable metric for quantifying evolutionary rates of clades across the tree of life. The EPI score is obtained by combining three elements: time (age of the clade in millions of years), success (relative success in terms of numbers of species), and change (relative number of changes since the clade first appeared). Since the last element is difficult to quantify, we follow the authors’ suggestion to use a simplified version of the metric (pEPI), which considers only time and success and is available for a broader range of clades. Furthermore, since pEPI values are mostly negative, we consider for convenience the opposite values, which we term opEPI, obtained by multiplying pEPI values by -1. Species with higher opEPI values are characterised by a higher degree of “living-fossilness”, or higher evolutionary stasis.

To test ESTA’s prediction, we compared the longevity values of some living fossils (target species) with that of other species (contrasting species). The targets are taken from Table 1 of [6], which reports the top 5 living fossils for 9 categories of animals. To allow for a meaningful comparison, each target species should be contrasted with a related species in the biological taxonomy, ideally one that shares as many taxonomic ranks (i.e., class, order, family, genus) as possible with the target. To control for the well-documented positive correlation between body size and longevity [38], each pair was matched for approximate body weight. A persistent challenge was identifying the appropriate taxonomic resolution for comparison, as the association of living fossils with a specific taxonomy level is not straightforward.

opEPI scores are sourced from Appendix 6 of [6], which provides a comprehensive list of 21,742 clades together with their corresponding pEPI values. To complement these measures of evolutionary stasis, longevity, and body weight data were obtained primarily from the AnAge database [10], while taxonomic information was mainly derived from the NCBI taxonomy database [28]. These sources have been supplemented with additional data where necessary to address any gaps. while clades with substantial missing information were excluded from the analysis. The final dataset comprised six categories (amphibians, birds, lepidosaurs, mammals, vertebrates, and actinopterygians), with three representative species selected for each.

The table obtained (Fig. 3) reports opEPI, longevity, and weight for 18 target species and as many contrasting species. The first column (first common node) reports the lowest taxonomic level shared by the two species, and the second column (comparison node) reports the taxonomic level at which the comparison is done. As can be seen, the target species generally exhibit higher longevity values compared to the contrasting species, in agreement with ESTA’s predictions. The effect can be quantified by calculating for each target-contrasting pair the difference between their respective longevity values: these differences have a mean of 15.274 and a standard deviation of 20.645. It should be noted, however, that this relationship is not uniformly clear across the entire tree of life, as the selected pairs represent only a limited subset. Accordingly, the table should be read as an exploratory synthesis that captures directional trends rather than as a definitive survey.

Fig. 3.

Examples of living fossils (displayed with a white background) are juxtaposed with closely related (contrasting) species in the biological taxonomy (highlighted with a gray-shaded background). The first column (first common node) reports the lowest taxonomic level shared by the two species, while the second column (comparison node) reports the taxonomic level at which the comparison is done. The letters indicate the taxonomic level (P: philum; C: class; O: order; F: family; G: genus). For each species, three parameters are reported: opEPI score, lifespan (in years), and weight (in kilograms)

Evo-devo dynamics

The idea that embryonic development reflects evolutionary history was formalised by Ernst Haeckel in the 19th century as the “biogenetic law”, commonly summarised as “ontogeny recapitulates phylogeny”. In his view, the developing embryo passes through stages that mirror the adult forms of its evolutionary ancestors, with new evolutionary modifications added at the end of development [20]. This bold claim became highly influential in early evolutionary biology and comparative embryology, shaping the belief that development provided a direct window into phylogeny. However, Haeckel’s model relied on oversimplified comparisons and, in some cases, inaccurate illustrations, leading to growing criticism even among his contemporaries [33]. This section does not revive the biogenetic law in a literal sense, but draws on its historically influential intuition within a modern evo-devo framework.

Although Haeckel’s original formulation is no longer accepted, modern biology has revived a weaker, more biologically grounded interpretation of the relationship between development and evolution. Contemporary research recognises that early developmental stages tend to be highly conserved across species, whereas evolutionary innovations often appear in later developmental phases, an updated concept sometimes referred to as “terminal addition” [19] and “peramorphosis” [1]. Rather than literal recapitulation, development is now understood as reflecting deep homologies, modularity, and hierarchical gene regulatory networks that shape how evolutionary changes manifest during ontogeny. Thus, while the biogenetic law is obsolete in its strict form, the intuition that development retains a detectable imprint of evolutionary history remains present in modern evolutionary developmental biology.

The features of conservation of early developmental stages and terminal addition were already present in an earlier formulation of the ET model, developed before ESTA was introduced. In [17], several scenarios were compared to evaluate evo-devo performance under different parameter settings, without reference to ageing. In the “progressive freezing” scenario, developmental stages are evolved in sequence: each stage is assigned a fixed number of generations to evolve, after which it becomes “frozen” (i.e. no further modifications are allowed) and its developmental genes are inherited unchanged by all individuals in the population. Formally, this scenario corresponds to setting the mutation probability to zero for developmental genes in all but the last, currently evolving stage, yielding a strict implementation of Haeckel’s principle.

In the simulations reported in [14], a more refined mechanism was introduced: the mutation probability of developmental genes in the genome was made dependent on the associated gene timers. In this formulation, the higher the timer value, the greater the probability of mutation. Simulations incorporating this timer-dependent mutation probability exhibited markedly superior evolutionary performance. Furthermore, [15] tested the possibility of evolvable timers, in which change events occurring later in life could gradually be shifted to earlier stages through adjustments in the associated timer values. This prediction was confirmed by the simulations, demonstrating the spontaneous emergence of earlier deployment of advantageous late-acting events.

These results suggest a way to reconcile Haeckel’s original insight with modern evo-devo principles. If the ESTA hypothesis is correct, and mutation probability increases with the timer value, then the latest developmental stages are those most exposed to ongoing selection, consistent with the phenomenon of terminal addition. The progressive freezing process explains also the similarities observed among embryos of different species (often most pronounced in the earliest stages of development). As lineages diverge, speciation events often recruit innovations late in ontogeny, reflecting each lineage’s evolutionary history while leaving the early developmental program more conserved across related species.

At the same time, the ability of evolution to shift the timing of change events (by modifying gene timers) blurs the boundaries between developmental stages and yields a more flexible, biologically realistic picture. Such timing shifts can also propagate downstream changes into earlier phases of development, softening the strict separation between conserved early stages and variable late stages. Together, these dynamics render the phenomena of terminal addition and early-stage conservation more nuanced and biologically plausible. Notably, the evolvability of developmental genes’ regulatory parts (defining target cells and timers) naturally incorporates the ideas of heterotopy (spatial shifts) and heterochrony (timing shifts), which are part of modern evo-devo theory.

Lifespan and sexual maturation

The third ESTA prediction concerns a corollary of the phenomenon of terminal addition. Groups of individuals or species exposed to stronger competition undergo faster evolution and, consequently, faster ageing. This process has two major outcomes. First, the post-reproductive decline becomes steeper, because a greater number of late-acting developmental genes are recruited, increasing the frequency of failed somatic “experiments.” Second, accelerated evolution introduces additional optimised stages at the end of development (terminal addition). As a result, the age of sexual maturation, which marks the end of development, is progressively advanced in the course of generations (scenario C in Fig. 4).

Fig. 4.

Life can be viewed as unfolding in two phases: development, driven by developmental genes already optimised by past evolution, and ageing, driven by genes still undergoing optimisation by present evolution. The timers associated with developmental genes are reflected in their position along the developmental time axis, and their degree of optimisation is represented by the green shading (darker indicates more optimised). Scenario A (baseline competition): many late-acting genes remain active, producing failed outcomes that reduce performance. Scenario C (higher competition): additional developmental stages are appended to the end of development, resulting in an advancement of the age of reproduction

An illustrative example of this shift comes from comparisons between two primate species, which should be viewed as suggestive rather than definitive, given the limited number of species available for analysis. Bonobos (Pan paniscus) and chimpanzees (Pan troglodytes) are two closely related species which diverged around 0.86–0.89 million years ago [48], with chimpanzees living north and bonobos south of the Congo River. The northern habitats, where chimpanzees live, are primarily characterised by dense forests and a more variable climate, and often experience greater competition for resources and harsher weather conditions. In contrast, the southern regions, where bonobos are found, tend to have a more stable climate and a more abundant food supply. Chimpanzees live in more hierarchically structured and often more violent societies, facing higher levels of intra-species conflict. Bonobos are known for their more peaceful and cooperative societies and experience lower levels of social stress.

Bonobos and chimpanzees exhibit broadly similar adult survival profiles and overlapping maximum longevities, both in the wild and in captivity [22, 43]. Despite this similarity, their life-history schedules differ in notable ways. In particular, bonobos tend to reach sexual maturation earlier than chimpanzees: females typically experience menarche around 6–8 years in bonobos, compared with 8–10 years in chimpanzees [5]. These observations are consistent with an association between the stronger evolutionary pressure experienced by chimpanzees and a corresponding shift toward later sexual maturation, consistent with ESTA’s prediction.

Another indication of the shift of sexual maturation comes from sex-specific life history differences across taxa. In mammals, females generally live longer than males, a pattern observed in both wild and captive populations [39]. In humans, women consistently outlive men across populations and historical periods [4], a phenomenon often discussed in the context of the grandmother hypothesis [21]. These lifespan differences are often accompanied by sex-specific differences in the timing of sexual maturation: in many species, including humans, females typically reach sexual maturity earlier than males [32]. Besides genetic determinants, these patterns may also be shaped by social, behavioural, and environmental influences.

In many animal species, males experience stronger competition than females, a phenomenon widely recognised under sexual selection [2]. This competition among males often takes the form of physical contests, dominance fights, or competitive displays, while females invest more in choosiness and selection. As a result, life-history costs for males tend to be higher: larger body size, aggressive behaviour, repeated fights, and high investment in competitive traits. Interestingly, across mammals (including humans) the rate of germline mutation is substantially higher in males than in females, with the paternal-to-maternal mutation ratio estimated to reach 4:1 in some species [11].

Together, these patterns are broadly consistent with the predictions of ESTA. Across evolutionary history, males have typically experienced stronger reproductive competition, a pressure that may favour the recruitment of additional late-acting developmental genes that contribute to a steeper post-reproductive decline and increased ageing rates. In parallel, late-acting developmental genes undergo optimisation and contribute to the emergence of additional stages appended to the end of development (with the nuances discussed in the previous section), resulting in a shift in the age of sexual maturation. Intraspecific sex differences are expected to be attenuated relative to interspecific contrasts, since males and females share the same genome and both sexes can benefit from advantageous innovations irrespective of their origin. At the same time, it should be acknowledged that multiple interacting factors (genetic, ecological, and social) are likely to contribute to the shaping of life-history traits.

Discussion

The relationship between environmental competition, evolution, and ageing is complex. SEAT predicts that high extrinsic mortality in harsh environments leads to shorter lifespans [34], because natural selection places less weight on optimising life stages that individuals are unlikely to reach. In contrast, in environments where individuals survive longer, selection can act more effectively on late-life traits, resulting in extended and improved later-life functioning. The perspective offered by ESTA contrasts with SEAT. Under this framework, intense environmental competition increases the rate of evolutionary change, leading to a greater number of evolutionary trials and, consequently, more frequent late-life dysfunctions. In less demanding environments, reduced selective pressure allows developmental programs to stabilise, resulting in slower ageing.

Table 1 presents the plausible causal sequence of events for two scenarios (“Easy and stable” and “Hard or unstable” environments) for both SEAT and ESTA. As we can see, for both SEAT and ESTA, low competition levels are associated with slow ageing rates. However, while SEAT predicts that this is achieved through a strong evolutionary activity, for ESTA, the opposite holds true. While SEAT associates increased evolution with less competitive environments (where it is needed to improve body maintenance at older age), ESTA links increased evolution with more competitive environments (where it is needed to help individuals to survive). In less competitive environments, ESTA foresees reduced evolution rates. The results of our analysis offer preliminary support for this concept, though their interpretation is not straightforward, and further data will be needed to assess its robustness.

Table 1.

Environment scenarios for SEAT and ESTA

	Environment easy and stable		Environment hard or unstable
	SEAT	ESTA	SEAT	ESTA
(1)	Weaker competition	Weaker competition	Stronger competition	Stronger competition
(2)	Longer survival	Slower evolution	Shorter survival	Faster evolution
(3)	Faster evolution	Slower ageing	Slower evolution	Faster ageing
(4)	Slower ageing	Longer lifespan	Faster ageing	Shorter lifespan

The prediction of a shift in the age of sexual maturation runs counter to the classical SEAT framework. ESTA predicts that under conditions of intense competition and high extrinsic mortality, evolutionary turnover accelerates, leading to faster ageing and later sexual maturation. This process is thought to unfold over evolutionary timescales. By contrast, SEAT predicts faster ageing in these same environments for a different reason: when survival is uncertain, natural selection reduces investment in somatic maintenance and repair, favouring a “live fast, die young” strategy supported by increased reproductive effort and earlier sexual maturation [40].

These two mechanisms need not be mutually exclusive. They can coexist if ESTA processes are understood as acting primarily through developmental genes (likely involving non-coding regulatory elements) while SEAT processes are hypothesised to act mainly through protein-coding genes that govern resource allocation, fertility, and somatic maintenance. Under this view, ESTA operates over longer evolutionary timescales, gradually reshaping developmental programs, whereas SEAT acts over shorter timescales, enabling rapid adaptive responses to ecological pressures.

ESTA offers a natural explanation for why evolution tolerates a seemingly detrimental phenomenon such as ageing. At first glance, individuals without ageing would produce more offspring and should therefore be favoured by natural selection. However, the response to this is straightforward: evolution could disable ageing, but doing so would effectively disable itself. By eliminating ageing, the evolutionary system would lose its primary mechanism for continuous adaptation and refinement. This strategy might only be advantageous in the absence of competition (whether inter-species, intra-species, or environmental), a scenario that is highly unrealistic in natural ecosystems. With few exceptions, the lineages that may have attempted this path are not around to tell the tale.

Conclusions and future work

In this work, we present observations that are consistent with several key predictions of ESTA. First, the theory proposes a link between evolutionary dynamics and rates of ageing. We explored this idea by examining species occupying ecological niches associated with reduced competition, as well as taxa commonly described as living fossils; in both cases, extended lifespans are frequently observed. Second, the framework offers a modern reinterpretation of classical recapitulation ideas, combining the principle of terminal addition with contemporary evo–devo concepts such as heterochrony and heterotopy. Third, the theory suggests that stronger evolutionary pressures may be associated with shifts in the age of sexual maturation. This expectation finds preliminary support in comparative patterns observed between bonobos and chimpanzees, as well as in well-documented sex differences in maturation timing across many mammalian species.

It must be emphasised that biological data are inherently open to multiple interpretations, and no single framework can claim to provide a definitive account of complex processes such as ageing. The primary objective of this paper has been to demonstrate that ESTA is not only consistent with empirical observations but also offers a complementary perspective to existing theories of ageing. In this context, the results presented here represent an initial step toward a more rigorous evaluation of whether ESTA may ultimately provide a broader, more integrative, and unifying explanation of ageing and evolution.

The road ahead presents several promising directions for further testing ESTA. One working hypothesis is that developmental genes are encoded predominantly in lncRNAs, rather than in protein-coding genes. Future work will therefore focus on detecting evo-devo signals specifically within this genomic category. Under ESTA, genetic regions influencing late-life events are expected to be comparatively less constrained and more variable, allowing selection to act on their phenotypic effects over time. In this framework, late-acting alleles are not interpreted as evolutionary remnants but as variants that remain subject to ongoing evolutionary refinement. If this prediction holds, signatures of elevated variability and lineage-specific divergence should be detectable within lncRNA loci involved in late developmental regulation.

At the same time, many protein-coding genes are known to influence lifespan across species, and interventions such as calorie restriction demonstrably modulate ageing through well-characterised genetic pathways. These observations indicate that ageing may be shaped by more than a single process, and disentangling the contributions of these distinct mechanisms will require further dedicated studies. Within this framework, one possibility is that long-lived species may undergo relatively rapid evolutionary change in protein-coding genes, while simultaneously exhibiting evolutionary stasis in lncRNA elements that encode developmental instructions.

Author contribution

Not applicable (there is only one author).

Funding

No funding was received for conducting this study.

Data Availability

Not applicable.

Declarations

Ethics approval and consent to participate

Not applicable.

Competing interests

The author declares no competing interests.