A substantial body of research has tested the hypothesis that exposure to adverse rearing conditions and cues of extrinsic mortality (EM) risk can accelerate maturation (Belsky, Steinberg, & Draper, Reference Belsky, Steinberg and Draper1991; Draper & Harpending, Reference Draper and Harpending1982) and initiate other components of a “fast” life history characterized by earlier age at first birth, shorter interbirth intervals, and greater completed fertility (Chisholm, Reference Chisholm1993; Ellis et al., Reference Ellis, Figueredo, Brumbach and Schlomer2009). Maturational acceleration in relation to EM has most consistently been reported in places like the US, where adversity frequently co-occurs with the maturation-accelerating effects of overweight. In contrast, there is little evidence that EM cues accelerate puberty in poorer quality energetic environments (Kyweluk et al., Reference Kyweluk, Georgiev, Borja, Gettler and Kuzawa2018), which slow growth rate and thus delay maturity (Eveleth & Tanner, 1976, Reference Eveleth and Tanner1990). In contrast, there is more consistent cross-cultural evidence that EM cues can accelerate behavioral components of life histories, such as earlier age of sexual debut (Gettler et al., Reference Gettler, McDade, Bragg, Feranil and Kuzawa2015). Based on these findings, and building on prior similar ideas (Coall & Chisholm, Reference Coall and Chisholm2003), we have argued that growth rate and pubertal timing are primarily driven by energetics, while reproductive behaviors are relatively uncoupled from energetics and thus free to vary independent of energetic conditions (Gettler et al., Reference Gettler, McDade, Bragg, Feranil and Kuzawa2015; Kyweluk et al., Reference Kyweluk, Georgiev, Borja, Gettler and Kuzawa2018).

Ellis and colleagues articulate a largely similar model but propose that life history decisions are coordinated by a 2-tiered regulatory system. It is worth noting that the three co-authors of this commentary had difficulty agreeing on what was implied biologically by “tiers” (e.g., does it imply a threshold effect, or additive and opposing effects of nutritional stress and EM?). We mention this because this may point to unnecessary complexity in the model being proposed. We feel that the idea that pubertal timing is energy-limited, and behavioral components of life history are not, provides a simpler and arguably more parsimonious interpretation of most of the findings reviewed in this target article.

It is first important to note that most of the evidence that the authors review is agnostic to the two-tier model, because most only demonstrates relationships between EM cues and reproductive behaviors. That fertility is higher in high-mortality settings has been well-known for roughly a century (Notestein, Reference Notestein and Schultz1945). The prediction of accelerated maturation due to EM cues is more debatable because (as noted) it runs counter to extensive evidence that adverse environments slow growth and delay puberty (Eveleth & Tanner, 1976, Reference Eveleth and Tanner1990). Ellis and colleagues review evidence to support this idea in the form of accelerated puberty in marginally nourished but high-mortality smaller-scale societies (Walker et al., Reference Walker, Gurven, Hill, Migliano, Chagnon, De Souza, Djurovic, Hames, Hurtado, Kaplan, Kramer, Oliver, Valeggia and Yamauchi2006). Although these data are notable, we agree with the authors that they are challenging to interpret. In this small sample of populations and ecological settings, accelerated puberty likely partially reflects genetic selection. Some of the earliest-maturing populations included in the review are also among the world’s shortest; in these groups, short stature per se may be the target of selection (Venkataraman et al., Reference Venkataraman, Yegian, Wallace, Holowka, Tacey, Gurven and Kraft2018) and a result of convergent genetic adaptation to long-term habitation of rainforest environments (Bergey et al., Reference Bergey, Lopez, Harrison, Patin, Cohen, Quintana-Murci, Barreiro and Perry2018).

Even if plasticity is a contributor, it should not be assumed that this response reflects an adaptation (Gould & Lewontin, Reference Gould and Lewontin1979). When a trait has a deep phylogenetic history, such as the impact of glucocorticoids like cortisol on maturational tempo across vertebrates (Crespi & Denver, Reference Crespi and Denver2005), current expressions of the trait may partly reflect vestigial effects that are no longer adaptive (Kuzawa, Reference Kuzawa2005). As we have argued (Kyweluk et al., Reference Kyweluk, Georgiev, Borja, Gettler and Kuzawa2018) for selection to generate the ability to accelerate puberty in response to EM cues, hominin ancestors must have routinely faced (1) high EM and (2) an abundant energetic environment simultaneously. Meeting the first requirement, adult EM risk – typically driven by disease, accidents, and violence – is relatively high in tropical/warm climate forager societies, approximating those in preindustrial European settings (Gurven & Kaplan, Reference Gurven and Kaplan2007). Regarding the second condition, forager children’s age at menarche is tied to nutritional conditions (Hochberg, Gawlik, & Walker, Reference Hochberg, Gawlik and Walker2011), and their nutritional status is often marginal and with high rates of stunting (Blackwell et al., Reference Blackwell, Pryor, Pozo, Tiwia and Sugiyama2009). This can result from food insecurity as well as infection, including malnutrition secondary to the symptoms of illness (diarrhea, anorexia), and the energetic costs of immune activation (McDade et al., Reference McDade, Reyes-Garcia, Tanner, Huanca and Leonard2008; Urlacher et al., Reference Urlacher, Ellison, Sugiyama, Pontzer, Eick, Liebert, Cepon-Robins, Gildner and Snodgrass2018). We find it unlikely that high EM would have co-occurred with abundant nutrition often enough to favor a contingent capacity for accelerated puberty (Kyweluk et al., Reference Kyweluk, Georgiev, Borja, Gettler and Kuzawa2018).

One area that Ellis and colleagues’ model differs from ours is in emphasizing that energetics can constrain fecundity. We agree that this is important to consider: cues may motivate behaviors that accelerate reproduction, but nutritional stress could suppress the underlying physiological processes (e.g., sexual behavior will only lead to pregnancy if there is sufficient energy available for ovulation). Although the probability of conception and resumption of menses are sensitive to energetics (Ellison, Reference Ellison2001), populations in which measures of nutritional stress are highest (e.g., infant growth faltering, adult stunting, rates of underweight) also tend to have the highest fertility rates (Bongaarts, Reference Bongaarts1980). As nutrition improves and mortality declines in a population, fertility typically goes down. Thus, outside the context of severe food shortages (where it clearly matters), the degree to which nutritional status explains variation in human fertility or population fertility trends is not clear. Nonetheless, we agree that nutrition not only has effects on growth rate and thus age at maturity but also has the potential to place physiologic limits on fecundity and fertility.

European history has been marked by dramatic changes in nutrition, growth, maturational tempo, mortality, and fertility, providing useful insights into how these traits tend to covary through time. In Western Europe and Scandinavia, pubertal onset declined gradually from as late as 17 or 18 down to 12 years since the mid-1800s (Wyshak & Frisch, Reference Wyshak and Frisch1982), which occurred because better-nourished children grow faster (Eveleth & Tanner, Reference Eveleth and Tanner1990). This 5–6 year acceleration greatly exceeds any documented acceleration predicted by EM cues and occurred as mortality rates declined. And as noted, although energy constraints can impair fecundity, populations with the lowest development indices, where food security is lower and undernutrition more common (Soheylizad et al., Reference Soheylizad, Ayubi, Mansori, Gholamaliee, Sani, Khazaei, Hanis, Shadmani and Khazaei2016), tend to have the highest fertility rates (Furuoka, Reference Furuoka2009). We feel that these and other past findings support our interpretation: that energetics primarily constrain growth (and thus maturational tempo), while life history-relevant behaviors, including reproductive scheduling, can vary independently of nutrition in response to EM cues (along with even more important cultural/socioeconomic drivers). Although we applaud Ellis and colleagues for stimulating discussion of this important literature, we do not see the value of reframing these findings in the language of tiers, which are mechanistically ambiguous and introduce unnecessary complexities.