Keywords

In evolutionary terms, humans have a unique life history strategy compared to other primates, including our closest relatives, the Great Apes. In particular, we have highly vulnerable infants who grow and mature very slowly, often remaining reliant on investments from caregivers well into their second decade of life (Hrdy, 2009; Kaplan, Hill, Lancaster, & Hurtado, 2000; Kramer, 2010). This period of protracted and slow physical growth for our children is physiologically linked to the very high metabolic costs of our large brains (for body size), which is a distinguishing characteristic of our evolution as a species (Charnov & Berrigan, 1993; Kuzawa et al., 2014). Despite these elevated time- and resourced-based costs of raising our children to independence, humans often “stack” the dependency periods of offspring on top of one another, raising multiple costly offspring simultaneously, rather than waiting for each one to mature to independence before having another (as is common in almost all other primates) (Kaplan et al., 2000; Kramer, 2010). Consequently, in contemporary natural fertility populations, humans have substantially shorter spaces (inter-birth intervals) between children than is common in our closest primate relatives. Despite this “stacking” of the intensive costs of our vulnerable children, comparative demographic research in small-scale societies, such as foragers or hunter-gatherers, has shown that our offspring also survive to adulthood at much higher rates than do Great Ape young (Kaplan et al., 2000).

Hence, humans have evolved a suite of reproductive and developmental characteristics that cross-species comparative perspectives would suggest should come with trade-offs (e.g., producing more young in shorter periods of time would typically be traded off against offspring survival rates) (Kaplan et al., 2000; Kramer, 2010). In the evolutionary past, hominin mothers would not have been able to manage the energetic costs of raising multiple costly human/hominin offspring simultaneously. A female employing such a strategy (alone) would have experienced lower reproductive fitness (i.e., likely as a result of greater offspring mortality) relative to her peers who continued to engage in a reproductive strategy involving raising one offspring at a time, as is found among the Great Apes (Kaplan et al., 2000; Kramer, 2010). Consequently, in scholarship and frameworks focused on human evolution, it is virtually uniformly held that in ancestral human/hominin populations, mothers would have needed extensive help from cooperative caregivers (e.g., grandmothers, fathers, older siblings, non-parental kin, etc.) in order for the life history strategy I have just described to have emerged evolutionarily in our lineage (Hawkes & Coxworth, 2013; Gettler, 2010; Gray & Crittenden, 2014; Kramer, 2010). Within this framework, often referred to as “cooperative breeding,” humans have specifically evolved the capacity for committed, involved fathering (Gettler, 2014; Gray & Anderson, 2010). If this represents an adaptive suite of behavioral tendencies and capacities, then evolutionary theoretical perspectives, which I will describe below, posit that human males should have hormonal and neurobiological mechanisms that help facilitate the expression of that commitment (Gettler, 2014; Rosenbaum & Gettler, 2018). The study of those mechanisms has grown into an area of research that is sometimes referred to as “the biology of fatherhood,” with a large body of literature on men’s testosterone (T), a growing body of work on fathers’ oxytocin, and comparatively less research on other pertinent psychobiological mechanisms such as prolactin, vasopressin, and cortisol (Abraham & Feldman, 2018; Gettler, 2014; Gray, McHale, & Carre, 2017; Rosenbaum & Gettler, 2018).

In this chapter, I first review evolutionary and (cross-species) comparative perspectives that serve as core foundations for making predictions about human fathers’ psychobiology. Then, I review research on men’s T, partnering/parenting status, and parenting behavior as an exemplar for the relevance of evolutionary perspectives to current understandings of contemporary men’s biology and family systems. While T is only one neuroendocrine mechanism relevant to fathering (Abraham & Feldman, 2018; Gettler, 2014; Rosenbaum & Gettler, 2018), it is the most widely studied and thus a review the literature on this hormone can be considered illustrative of broader principles and questions that are relevant to “the biology of fatherhood” more broadly.

Evolutionary Theoretical Perspectives on Human Fathering

Evolutionary perspectives, particularly parental investment theory and life history theory (LHT), are commonly used as framing lenses for research on the physiological underpinnings of fathering among humans and other animals (Fernandez-Duque, Valeggia, & Mendoza, 2009; Gettler, 2014; Gray et al., 2017). Compared to parental investment theory, LHT is a more expansive theoretical framework and is extensively used in evolutionary biology and ecology. It focuses on the ways in which organisms must allocate limited resources, particularly energy, to mutually exclusive physiological demands related to growth, reproduction, and survival, and how species have evolved to “solve” these allocation trade-off challenges through a range of “LH strategies” (Hill & Kaplan, 1999; Stearns, 1992). Across species, variations in these LH strategies, which result from organisms exploiting different ecological niches and thus experiencing diverse selective pressures (e.g., predation; infectious disease) through (deep) evolutionary time, are often viewed on a fast-to-slow continuum and include a focus on the timing-length of key life events (e.g., age at reproductive maturity; length of the reproductive window; total life span) and related physical-behavioral characteristics (e.g., adult body size; number of lifetime reproductive events; number of offspring per reproductive event; time-energy devoted to parental investment) (Charnov & Berrigan, 1993; Promislow & Harvey, 1990).

Physiological signals (such as hormones and neurotransmitters) mechanistically mediate the trade-offs between these various time and energy demands. Thus, genetically underpinned physiological profiles that differ between individuals become more common in the population (i.e., positive selection within-species) across generations, as some organisms achieve higher fitness and pass along their genes at greater rates than others (Bribiescas & Ellison, 2008; Stearns, 1992). Although organisms are limited by phylogenetic inertia (meaning, in this case, that their ability to developmentally acclimate is constrained by the evolutionary history of their species), it is expected that organisms will have some adaptive capacity to flexibly adjust their individual allocation strategies in relation to their current circumstances, which is related to the biological concept of reaction norms (Stearns, 1992). For example, a developing (younger) organism will tend to grow faster and bigger under conditions of energetic abundance compared to contexts with nutritional constraints, while an adult might reduce energetic investments in reproduction when faced with an environment with high levels of infectious disease, enabling survival despite lower immediate reproductive prospects (Hill & Kaplan, 1999).

These facultative (environmentally sensitive) within-individual LH shifts and within-species ranges of variation in LH strategies manifest themselves through physiological pathways (e.g., hormonal axes and neurobiological systems). This multi-level perspective on variation serves as a foundation for considering individual-level differences in LH-relevant biological systems and their emergence or calibration across the life course in response to environmental conditions (Bribiescas & Ellison, 2008; Jasienska, 2013). In terms of fathering and its underlying biology, LHT provides a predictive framework for the ways in which adults (particularly male mammals, see below) face core trade-offs between mating and parenting in their allocations to reproductive effort when biparental care evolves (Fernandez-Duque et al., 2009; Gettler, 2014; Hill & Kaplan, 1999). While recent psychobiological models have provided a substantive, nuanced perspective challenging the applicability of a dichotomous breakdown of mating vs. parenting effort to human neuroendocrine physiology and behavior (van Anders, 2013; van Anders, Goldey, & Kuo, 2011), a substantial amount of work on the psychobiology of human fatherhood, particularly for T, has drawn from the LHT perspective to make predictions about the ways in which men’s hormonal production will vary based on the transition to parenthood (and away from mating effort) or men’s specific investments in parenting effort (e.g., direct caregiving) and cooperation with partners to raise children (Gettler, 2014; Gray, McHale, & Carre, 2017).

In that vein, a substantial amount of the existing research on the biology of fatherhood (across taxa) has focused on variability in testosterone (T) based on the Challenge Hypothesis, an evolutionary biological model that emerged from the study of birds and is used to explain LH trade-offs between mating and parenting (Gray, Straftis, Bird, McHale, & Zilioli, 2019; Wingfield, Hegner, Ball, & Duffy, 1990). In short, drawing on extensive ornithological behavioral physiological research, Wingfield et al. (1990) proposed that among species in which males invested time and energy in raising their young, their T would decline during the stages of the breeding season in which their offspring were dependent. This downregulation of T was argued to be a pathway that would facilitate males’ focus on cooperative parenting demands and shift their devotion of time/energy away from mating opportunities with females and competition with males for territory while also allowing males to attenuate the physical and metabolic costs of long-term T upregulation. Meanwhile, among species in which males had not evolved to cooperate with mothers and/or to invest in their young, they would not show this characteristic T decline, on average, at the “investment” stage of the reproductive cycle and their T would remain elevated across the breeding season if they were continually engaging in extensive male-male competition. It is important to note that it is very common for bird fathers to invest in their young, with up to 90% of bird species showing biparental care, which is a much higher rate than in other vertebrate taxa, particularly mammals (see below) (Clutton-Brock, 1991). Thus, in birds exhibiting biparental care, this perspective suggests that in many species males have evolved (via natural selection) the capacity to reduce their T to help shift their behavioral and energetic allocations toward priorities that help to optimize their cooperation with females and the promotion of the survival, growth, and development of their offspring (Wingfield et al., 1990).

Compared to the patterns among birds, paternal investment is rare among mammals, occurring in only around 3–5% of mammalian species (Clutton-Brock, 1991; Gray & Anderson, 2010). This low prevalence of mammalian paternal care almost certainly reflects divergences in male and female reproductive strategies that emerged alongside the evolution of internal fertilization, internal gestation and viviparity, and lactation that characterize mammalian reproduction (Clutton-Brock, 1991). Collectively, these characteristics often reduce males’ certainty of paternity and their ability to remain in proximity to their young between fertilization and eventual birth, contributing to the predominant mammalian pattern of males competing with one another for mating opportunities with multiple females, rather than partnering with a pregnant female and raising his (potential) offspring (Royle, Smiseth, & Kolliker, 2012). Because the hypothalamic-pituitary-gonadal axis that produces T is evolutionarily ancient and shared across vertebrates, the Challenge Hypothesis laid the groundwork for scientists to make predictions regarding shifts in T production based on mating and parenting effort in other vertebrate taxa in which biparental care evolved, including its rare occurrence in mammalian species, such as humans (Gray et al., 2019; Wingfield et al., 1990). Notably, any similarities in the physiology underlying paternal involvement among birds and mammals are the result of convergent evolution (rather than shared ancestry) and represent natural selection (or other evolutionary processes) repeatedly coopting similar neurobiological and hormonal pathways to promote paternal care, however diverse the ecological settings in which it evolves. This evolutionarily grounded approach to the study of the biology of vertebrate fathering has led to the observation that pair-bonded and/or invested vertebrate fathers commonly show similar physiological profiles, including particularly lower T during the periods in which they cooperate with mothers to raise vulnerable offspring (Fernandez-Duque et al., 2009; Gettler, 2014; Gray, McHale, & Carre, 2017).

Thus, this comparative perspective has served as a pillar for evolutionarily grounded biosocial or biocultural conceptual frameworks and empirical research aimed toward testing whether human fathers express psychobiological profiles similar to other invested vertebrate fathers and how those profiles might be shaped by the cultural and family system contexts in which they find expression (Abraham & Feldman, 2018; Gettler, 2014, 2016; Gray, McHale, & Carre, 2017; van Anders, 2013; van Anders et al., 2011). Human male capacities for intensive paternal care evolved alongside or downstream of the suite of life history characteristics I described at the outset of the chapter that distinguish humans from the Great Apes and other primates (Gray & Anderson, 2010; Kaplan et al., 2000). To briefly reiterate, human life history includes the birth of newborns who are relatively neurologically and physically underdeveloped, comparatively early weaning of those infants, the slow growth and development of our offspring through a unique “childhood” phase, and shortened inter-birth intervals (such that our hyper-dependent offspring are “stacked” on top of one another, under natural fertility conditions) (Kaplan et al., 2000; Kramer, 2010).

This suite of reproductive-fertility and developmental characteristics would have been too much of an energetic and time burden for a mother to successfully manage alone (in the evolutionary past), rendering it a maladaptive for reproductive fitness in the absence of the types of cooperative caregiving that we observe across contemporary societies (Kaplan et al., 2000; Kramer, 2010). From an evolutionary perspective, it is highly probable that fathers frequently made critical though variable contributions to the well-being of their children, along with mothers receiving assistance from other helpers, often referred to as allomaternal caregivers (their own older offspring, grandmothers, other female kin, etc.) (Gettler, 2010; Gray & Anderson, 2010; Hawkes & Coxworth, 2013; Kramer, 2010).

Conceptualizing Fathers’ Roles in the Evolutionary Past

There is general (though not universal) agreement that fathers’ provisioning of energetic resources was likely important to the evolution of human’s life history strategy by helping to improve child health and survival (Gray & Anderson, 2010; Gurven & Hill, 2009; Kaplan et al., 2000). In that vein, fathers’ contributions as providers have recently been linked to children’s improved nutritional status and growth in small-scale societies in which the environmental conditions include evolutionarily relevant features (i.e., energetic constraint; infectious disease stress) (Boyette, Lew-Levy, & Gettler, 2018; Winking & Koster, 2015). Compared to the importance of paternal provisioning, there has been less of a focus in evolutionary frameworks and related research on the roles of fathers as direct, physical caregivers and as contributors via other pathways (e.g., cultivators of social capital or status within the community) to child well-being (Boyette et al., 2018; Gettler, 2010; Scelza, 2010; von Rueden, Gurven, & Kaplan, 2011). Relative to fathers’ roles as providers, there are fewer lines of evidence available to reconstruct the potential roles fathers may have played as direct caregivers during our evolutionary past. In particular, in many contemporary societies, fathers’ roles in such care are relatively modest (Gettler, 2010), despite the increasing recognition that sensitive, warm, and supportive fathering can benefit child’s social and emotional development in settings such as the United States and Europe (Cabrera, 2020). Scholars working at the intersection of human psychobiology and social bonding have specifically argued that testing for physiological profiles that help facilitate nurturant, sensitive caregiving in contemporary fathers (and within other salient social relationships) can serve as one line of evidence to test whether such roles were evolutionarily adaptive, in light of cross-species comparative data and based on LHT (Gettler, 2014; van Anders, 2013).

In that vein, human fathers’ psychobiology has been repeatedly linked to engagement in direct caregiving and nurturance, although these neuroendocrine-behavioral profiles are by no means canalized or fixed (Gettler, 2014; Gray, Reece, et al., 2017; van Anders, 2013). The evolutionary perspective lays the foundation for understanding why (ultimately) human fathers have these psychobiological capacities, and provides a useful framework for making predictions regarding the types of behaviors and social contexts that are most likely to elicit (potentially) evolutionarily advantageous profiles. Yet, importantly, relatively recent psychobiological frameworks (van Anders, 2013; van Anders et al., 2011) have made the observation that “parenting” represents a diffuse range of demands that might be most optimally met by varied levels of (for example) T, rather than a singular profile of invested fathers having low T. For a number of reasons (including phylogenetic inertia), evolutionary processes result in traits that are “good enough” and not perfect; thus, human psychobiological responses to partnering/parenting are flexible but cannot be expected to reflect unlimited plasticity to respond to all aspects of family life and demands outside the family and might be occasionally mismatched to contemporary demands (Gettler, 2016). This provides an opportunity to build from LHT tenets as well as newer psychobiological frameworks (van Anders, 2013; van Anders et al., 2011) by integrating culturally-grounded, family system perspectives and by emphasizing individual differences that may shape the expression of the biology of fatherhood (Gettler, 2016). To help illustrate these points, in the next section, I will briefly describe some of the physiological functions of T that are considered relevant to life history trade-offs across vertebrate species. I will then build from that foundation to review the human literature on men’s T and partnering/parenting status. Finally, I will transition to a discussion of research exploring the relationships between men’s T and their caregiving behavior.

Testosterone (T), Partnering, and Parenting Status

As I described above, both LHT and parental investment theory focus on trade-offs that occur between core fitness-relevant demands, such as reproduction versus maintenance of the body, as a consequence of time and energy being limited resources. In particular, both theoretical perspectives are consistent with the notion that mammalian males will face reproductive effort trade-offs between dedicating limited time and energetic resources toward mating effort versus toward the cooperative care of young in those rare species exhibiting biparental care. Indeed, aspects of the Challenge Hypothesis, outlined above, are likewise premised on similar ideas regarding the ways in which male birds in species with biparental care will seasonally shift away from mating effort and toward cooperative parenting of young, as the two (mating and parenting) are often incongruent as to their physiological underpinnings and behavioral demands (Gray et al., 2019; Wingfield et al., 1990). Across vertebrates, T is considered a key physiological mechanism helping to mediate these mating versus parenting trade-offs. In vertebrate males, elevated T often facilitates behaviors related to competition for resources, territory, and status. Similarly, higher T generally facilitates costly investment in physical attributes, such as larger body size, musculature, and ornamentation (such as colored plumage in certain birds) that facilitate attraction of mates and competition for resources and status (Bribiescas, 2001; Gray, Reece, et al., 2017; Hau, 2007). Based on these patterns, many of the early studies of the biology of human fatherhood focused on men’s “life history status,” with the prediction that, in some settings, partnered fathers would have lower T than single non-fathers, particularly (Gray, Kahlenberg, Barrett, Lipson, & Ellison, 2002).

Over the past two decades, many studies have documented that partnered fathers tend to have lower T than single non-fathers, although, as I will discuss, this does vary by cultural context. Moreover, multiple thorough synthetic reviews have been written on this topic, which give far greater coverage of the relevant literature than I am able to do here (reviewed in Gettler, 2014; Gray, Reece, et al., 2017; Roney & Gettler, 2015; van Anders et al., 2011; van Anders, 2013). Importantly, a recent meta-analysis showed that there is an overall pattern of partnered men, especially fathers, having lower T than other men, though the prevailing effect size is relatively modest (Grebe et al. 2019). A subsequent meta-analysis that used a somewhat different approach in trying to isolate the relationship between T and fathering (adjusting for partnering) found largely complementary supportive results, but again with a relatively modest effect size (Meijer, van IJzendoorn, & Bakermans-Kranenburg, 2019).

My primary purpose here is to discuss some of the early work in this area, especially as it relates to cross-cultural patterns, and to highlight some of the larger and longitudinal studies that have been conducted on this topic, as these are less subject to research design limitations pertinent to smaller, cross-sectional studies. Gray and colleagues conducted a number of foundational studies in this domain, including work on men in the Boston area, showing that married men (regardless of parenting status) had lower T than unmarried non-fathers. Married fathers and married non-fathers did not differ for T in either analysis (Gray et al., 2002; Gray, Campbell, Marlowe, Lipson, & Ellison, 2004). In a similar study among Chinese men in Beijing, Gray et al. later found that married fathers had lower T than both unmarried and married non-fathers, hinting at some cross-cultural differences compared to the earlier studies in the United States (Gray, Yang, & Pope, 2006). Around this same period, Gray and colleagues also tested related ideas in two Kenyan cultural groups in which polygyny is culturally sanctioned and thus the dynamics of T-competition and mating-partnering/parenting are potentially blurred, such that men could potentially look more similar for T across life history statuses. Among Kenyan Swahili men, unmarried and monogamously married men did not significantly differ for T, while polygynously married men (i.e., with multiple partners) had higher T than other males (Gray, 2003). In contrast, in a larger study of Ariaal pastoralists in Tanzania, unmarried men in their reproductive primes had higher T than their married peers, while polygynously married men did not have elevated T relative to other men (Gray, Ellison, & Campbell, 2007). In a complementary study of Senegalese men later conducted by a separate set of researchers, married fathers similarly had lower T than unmarried non-fathers (Alvergne, Faurie, & Raymond, 2009). However, among fathers, polygynous men had higher T than monogamously partnered fathers, which is consistent with the idea that men’s mating effort and competition in the context of polygyny are correlated with elevated T (Alvergne et al., 2009).

These findings set helped to set the stage for a subsequent cross-cultural comparative study in Tanzania of two neighboring small-scale societies in which models of family life differ substantially (Muller, Marlowe, Bugumba, & Ellison, 2009). Among Hadza foragers, men and women generally partner monogamously and fathers frequently spend time in close proximity with their young children and are involved with hands-on care of them. In contrast, among the neighboring Datoga pastoralists, adult men spend much of their time away from women and young children, attending to responsibilities for herding and protecting livestock. Polygyny is also culturally sanctioned in this context (Muller et al., 2009). In their study, Muller et al. (2009) found that Hadza fathers had significantly lower T than non-fathers, whereas Datoga fathers and non-fathers had similar T. This was among the first studies to hint that father-child proximity and men’s direct caregiving potentially had implications for the patterning of men’s T based on parenting status, a theme I return to in the subsequent section on T and fathering behavior.

Subsequent longitudinal research built on these earlier studies to help address the “state” or “trait” question as it relates to T and men’s partnering and parenting (e.g., Edelstein et al., 2017; Gettler, McDade, Feranil, & Kuzawa, 2011). That is, because prior studies were cross-sectional, they could not address the question of whether men with reduced T were more likely to become partnered fathers in some settings (trait) or whether the transition to committed partnering and parenting led to declines in T (state). In research from a large, multi-decade birth cohort study in Metropolitan Cebu in the Philippines, my colleagues and I demonstrated that men who had higher T as single non-fathers in young adulthood were more likely to become partnered fathers by their mid-20s. Meanwhile, those men who became newly partnered new fathers between the ages of 21 (baseline) and 26 (follow-up) experienced large biologically meaningful declines in their T, with fathers of newborns showing the steepest declines (Gettler et al., 2011). These patterns differed significantly from the relatively static T profiles of men who remained single non-fathers over the same time frame. Other longitudinal research likewise found that fathers’ T was reduced in the first few months of parenthood, compared to non-parent controls (Perini, Ditzen, Hengartner, & Ehlert, 2012).

There are relatively few large, longitudinal studies of human psychobiology and partnering or parenting, but two such studies that focus solely on partnering complement our findings from the Philippines. In a large decade-long study of US military veterans, Mazur and Michalek (1998) showed that unwed men and men who changed marital status (married ↔ single) all had T that was consistently higher than men who remained married across the study period. Men who transitioned from being divorced to remarried showed significant declines in T during the decade of study. Meanwhile, married men who had elevated T at earlier time points in the study were more likely to become divorced by later time points. Drawing on the study in Cebu, our research team later showed a complementary finding showing that men with greater T functionality as young adults were more likely to experience relationship separation/dissolution by their mid-20s (Gettler et al., 2017). Additionally, in a recent, large, longitudinal study in Denmark, many of the core findings from Mazur and Michalek (1998) were replicated, as men who transitioned to marriage experienced declines in T and those that divorced/separated exhibited increases in T through time (Holmboe et al., 2017).

Finally, in addition to these longitudinal studies that help shed light on the bidirectional relationships between T and shifts in life history status, a small number of large cross-sectional studies have (in some cases) added clarity to these patterns, while also raising new questions. This is particularly important since it is known that studies with small sample sizes are prone to inflated effect sizes when they find statistically significant results (Button et al., 2013). In a US population representative study with over 1500 men ranging in age from 20 to 60 years old, my collaborator and I showed that partnered men residing with children had significantly lower T than men who were not residing with children and were either divorced or never married (Gettler & Oka, 2016). Men who were partnered but not residing with children tended to have modestly higher T than partnered men living with children but the effect size was small and did not reach statistical significance. These findings thus generally parallel the earlier foundation work on US men by Gray et al. (2002, 2004).

In contrast to these patterns that align with past work, data from a large study of military personnel (N = 4400+) found that for men in their early 30s, T increased based on the presence of children in their homes and increased as the number of children rose (Mazur, 2014). Based on the data reported in the study, it is not clear what explains the discrepancies between this study and other relevant work on US married fathers. However, contextualized within existing frameworks (Gettler, 2016; van Anders, 2013; van Anders et al., 2011), I discuss some possible explanations below in a subsequent section regarding the importance of models of fatherhood, cultural context, and individual differences. Finally, a recent work focusing on Jamaican fathers (n = 350) found that fathers’ T did not vary according to relationship status (married, long-term partnered but unmarried, visiting [mothers and fathers live apart but jointly raise children) (Gray, Reece, et al., 2017). This contrasts somewhat with my team’s earlier work from Cebuano men in the Philippines, which showed that residential fathers had lower T than non-residential fathers (Gettler, McDade, Agustin, Feranil, & Kuzawa, 2015). While overinterpreting non-significant findings is potentially fraught, these studies and the variation they point to within- and across-cultures points to the potential importance of local cultural models of fathering, partnering, and family life more broadly in influencing the expression of these psychobiological patterns (Gettler, 2016), which is a theme I focus on in the subsequent section.

Fathering Quantity in Relationship to T

As I have already noted, T is considered a key mechanism helping to facilitate trade-offs between mating and parenting effort. In terms of understanding how fathers’ T might affect and be affected by the functioning of family systems, there are a number of key points that merit discussion before delving into the relevant literature. First, the majority of studies that have drawn on LHT to formulate hypotheses about paternal care have generally focused on the quantity of caregiving (e.g., time spent in direct care) rather than the quality of care. Likely, because of its focus on time and energy as limited resources that are allocated to competing demands, LHT has not generally been used to generate predictions about caregiving quality, to date (Kuo & Gettler, 2018). In addition, there is a long-standing recognition that behaviors that are commonly lumped under the headings of “mating” and “parenting” effort, which are often a core focus of LHT approaches to modeling the trade-offs men negotiate when the transition to parenthood, are not necessarily mutually exclusive or divisible (Smuts & Gubernick, 1992; van Anders, 2013). For example, in many societies, men’s competition for resources and status could be considered relevant to both mating effort and parenting effort.

Nonetheless, a number of studies have found that when fathers are involved with direct care of their children and/or spend time in close proximity to them, they generally have lower T than fathers who are uninvolved in hands-on caregiving (Alvergne et al., 2009; Gettler et al., 2011; Gettler, McKenna, Agustin, McDade, & Kuzawa, 2012; Lawson et al., 2017; Mascaro, Hackett, & Rilling, 2013), although these findings are not ubiquitous, even in cultural settings in which they might be predicted to occur (e.g., Gray et al., 2002, 2004; see further relevant discussion below). A recent meta-analysis also found support for this pattern, but with a relatively small effect size, which (as the authors note) could potentially reflect limited attention to core contextual details regarding family systems and dynamics (Meijer, van IJzendoorn, & Bakermans-Kranenburg, 2019) as well as variability in study design (Kuo & Gettler, 2018). Nonetheless, a number of longitudinal studies have helped to shed some light on the direction of these effects. Research following US expectant fathers from the pre- to post-partum found that fathers whose T declined more across their partners’ pregnancies were more engaged in infant care post-partum and their partners reported that the lower T fathers helped more with household tasks and were more supportive (Edelstein et al., 2017). In a separate larger study of US fathers, Kuo and colleagues found that fathers’ whose T was lower the day after their babies were born reported participating in greater direct care and indirect parenting tasks related to the baby 2–4 months later (Kuo et al., 2018). Finally, my collaborators and I found that among Filipino men who were fathers at age 21–22 years, those who increased their time spent in childcare in the ensuing 4–5 years experienced declines in T over the same time frame, while fathers’ T went up if they decreased their caregiving time during the follow-up period (Gettler et al., 2015).

Much of this work is consistent with T playing a role as a mechanism to help shift men’s priorities as they transition to committed fatherhood, particularly in reference to a limited resource (i.e., time). This work is also broadly consistent with psychobiological frameworks pertaining to human social behavior and bonding that have helped to provide nuance to predictions regarding the ways in which men and women’s T will correlate with various demands related to partnering and parenting. Specifically, van Anders (2013) and colleagues (2011) have proposed that lower T will tend to correlate with nurturing behavior in the formation and maintenance of social bonds (such as can occur between partners or between parents and children), whereas elevated T will tend to be correlated with behaviors related to competition, including for status and resources. To the extent that nurturing parental behaviors (e.g., sensitive, warm, supportive interactions) are widely recognized as benefitting child development outcomes in the United States, Europe, and similar societal settings, this framework offers predictions for the way in which reduced T among fathers could be linked to child well-being in certain contexts (e.g., Landry, Smith, & Swank, 2006).

Fathering Quality in Relationship to T

Relatively few studies have tested whether lower T is related to the quality of fathers’ parenting and the findings in the existing literature are mixed (Kuo & Gettler, 2018). In studies of Israeli families, it was found that fathers who had lower T engaged in more affectionate touch and sensitive infant-directed speech (Weisman, Zagoory-Sharon, & Feldman, 2014) and were observed to be more behaviorally synchronized with their babies (Gordon, Pratt, Bergunde, Zagoory-Sharon, & Feldman, 2017). In an earlier work on Canadian men, fathers who had lower T expressed greater concern and sympathy toward a crying baby, when they were exposed to recorded infant distress (Fleming, Corter, Stallings, & Steiner, 2002). In the Netherlands, fathers’ waking and evening T, respectively, were not significantly linked to their sensitivity or respect for their children’s autonomy. However, fathers with steeper diurnal changes in T across the day were found to engage in higher quality parenting (as defined in the study) (Endendijk et al., 2016). The typical diurnal pattern reflects a peak around waking and nadir by early evening, which is reflected in the study (Endendijk et al., 2016). Thus, for the majority of men, the key finding of this research likely reflects that fathers with steeper diurnal declines in T across the day engaged in more sensitive, respectful parenting. Meanwhile, more recent work from the same research group has shown that fathers’ levels of self-control moderated the relationship between T and parenting quality. Specifically, low-self-control fathers with higher T were less sensitive and less respectful of child autonomy, whereas lower T in such fathers correlated with higher quality parenting. The reverse pattern was found for higher T fathers (van der Pol et al., 2019). This most recent (2019) research from this team points to the importance of bringing individual differences into the study of human paternal psychobiology as is done in other related domains such as research on T and aggression or physiological responses to psychosocially stressful conditions (Carré et al., 2017; Oswald et al., 2006).

Notably, however, a number of studies have failed to find significant correlations between fathers’ T and the quality of their parenting. For example, in a US-based sample, Kuo and colleagues (2015) did not find that fathers’ basal T was predictive of various measures of observed parenting quality, including sensitivity, responsiveness, and intrusiveness, during a lab-based father-infant teaching task. In a separate, relatively large US-based study, Dorius et al. (2012) found no significant link between fathers’ T and their children’s reports of father-child closeness. It is important to note that this study did not have direct measures of fathering quality. Theoretically, father-child closeness is predicted to emerge (in part) via high-quality (sensitive, warm, authoritative) parenting in this setting, thus, conceptually, T could plausibly have been associated with father-child closeness as a mediator in the pathway linking fathering quality and T. As Kuo and Gettler (2018) have recently described, it is difficult to discern whether inconsistencies in this literature reflect distinct psychobiological processes in different settings or samples, lack of statistical power (in some cases), or research design differences that lead to variation in what family and individual fathering dynamics are found to be meaningfully correlated with men’s T. For example, in the Dorius et al. (2012) study, the age range of children included was quite large (6–16 years old), and there are variable demands of parenting and emotionally connecting and bonding with children of those varying ages. Thus, “high-quality” parenting and “closeness” of parent-child bonds might well be correlated with differing psychobiological profiles in fathers as parenting demands shift with child age and development (Gettler, 2016; van Anders, 2013). In fact, the vast majority of the work I have reviewed above has been conducted with families with infants and toddlers in a relatively limited number of cultural settings. To better understand the bidirectional relationships between hormones such as T and men’s roles in families, we need to explore these questions across a more diverse landscape of family systems and societal settings (Gettler, 2016).

The models from van Anders and colleagues likewise point to the importance of social and cultural context in the expression of psychobiological systems, which is a critical consideration for interpreting the biology of fatherhood across diverse settings (van Anders, 2013; van Anders et al., 2011). Complementing those frameworks, I proposed a culturally oriented evolutionary developmental model (“Becoming DADS”), with the DADS acronym standing for Duration, Attitude, Dedication, and Salience. In this framework, I posited that evolutionarily conserved but developmentally plastic physiological systems, such as the hypothalamic-pituitary-gonadal axis that produces T, are calibrated to respond to and facilitate culturally valued social and gender roles within societies, based on boys’ childhood experiences within families and communities (Gettler, 2016). This model thus facilitates social neuroendocrine predictions based on locally valorized roles for and gender socialization of men as fathers and men’s individual experiences as children while also emphasizing the importance of broader processes that influence those dynamics, such as political economy and local ecology.

For example, drawing on my prior work on T and fatherhood among Filipino men (e.g., Gettler et al., 2011, 2015), I discussed how political economic changes in the Philippines in the 1980s (especially via neoliberalization of the economy) led to many Filipino women traveling overseas to seek better paying work and also a feminization of labor opportunities in Cebu, the Philippines. I argued that these shifting labor and migration dynamics potentially helped to contribute to changing family roles regarding childcare, requiring some fathers to participate more intensively in childcare, running counter to prevailing cultural norms of masculinity and expectations of fathers. In families and communities transformed by these changes, the developmental experiences of children would have been shifted (i.e., boys receiving some care from their fathers and thereby internalizing gendered roles and identities and developing bio-behaviorally in tune with those culturally salient dimensions of family life) (Gettler, 2016). My colleagues and I recently published work on the multi-generational patterning of fathering behavior and parenting identity in Cebu (the Philippines) that align with the general predictions of this framework, as boys who had fathers who did some caregiving in the early 1980s and developed close relationships with them were, themselves, more involved as fathers ~25 years later and described caregiving as more central to their identities as parents (Gettler, Kuo, Bas, & Borja, 2019). However, in that study, our core findings were based on crossover statistical interactions, meaning that for boys who grew up with fathers who were uninvolved with childcare and nonetheless developed a close relationship with them, they (the second-generation men) grew up to recapitulate their fathers’ parenting styles, with relatively less caregiving and less emphasis on it as a component of their parenting identities (Gettler et al., 2019).

Notably, our team observed these crossover effects for intergenerational patterning of paternal parenting styles in Cebu using data from the same sample on which we had previously shown that men becoming newly partnered new fathers had larger longitudinal declines in T than men remaining single non-fathers (Gettler et al., 2011). Our more recent (2019) results suggest that there was likely a meaningful subset of new fathers in that pool who did not see nurturant caregiving as a core part of their parental responsibilities or identities as fathers. Such individual-level insights might help to explain why, despite group-level aggregate findings such as declining T among newly partnered new fathers in Cebu (Gettler et al., 2011), we still see that ~25% of those new fathers exhibit relatively static or even increasing T across that same 4.5-year time period. A similar within-culture perspective on variation in models of fathering may help explain results I discussed previously from Mazur (2014) in which relatively young military fathers with more children had higher T, as opposed to lower levels that might be predicted based on increasing caregiving demands of more dependents as has been found in other research on non-military US fathers (Mascaro et al., 2013). In particular, there is a possibility that the orientation and engagement of military men in the sample to fathering were more consistent with competition (e.g., providing resources) and protection than nurturance, and those social neuroendocrine dynamics could explain the observed patterns (Mazur, 2014). While that interpretation is speculative, my colleagues and I recently published research on fathers’ roles and psychobiology from a small-scale society in the Republic of the Congo that gives it credence and that helps to lay the groundwork for the importance of understanding local models of fathering and family life vis-à-vis men’s neuroendocrine profiles.

This new research focuses on a small Bantu-speaking community of less than 200 who identify ethnically as Bondongo. They reside in a remote region of northern Republic of the Congo, far removed from major urban centers and with little market integration, and they subsist primarily via fishing and (non-mechanized) farming (Boyette et al., 2018). In this society, fathers’ roles as providers of resources are culturally valued and those who are seen as better providers have children in better health (Boyette et al., 2018). Fathers’ acquisition of resources often involves risky behaviors (e.g., scaling tall trees; fishing in dangerous waters) and men achieve higher status in the community through their abilities as providers. Meanwhile, the Bondongo do not culturally emphasize warm father-child relations or fathers’ direct caregiving, including sensitive or nurturant involvement, and such behaviors are relatively uncommon (Boyette, Lew-Levy, Sarma, & Gettler, 2019; Gettler, Sarma, et al., 2019).

Thus, in contrast to prior research on paternal psychobiology in settings such as the United States/Europe, the Philippines, and Israel, “higher quality” paternal care among the Bondongo is unlikely to be linked to lower T. In contrast, a culturally situated and theoretically grounded hypothesis would be that committed Bondongo fathers would exhibit higher T (Gettler, 2016; Gray, McHale, & Carre, 2017; van Anders et al., 2011; van Anders, 2013). We found support for this prediction, as fathers who were seen as better providers had higher T than men seen as less effective providers, which we suggest is consistent with theorized links between risk taking, competition for status, and elevated T (Gray, McHale, & Carre, 2017; van Anders et al., 2011; van Anders, 2013). To our knowledge, our findings were the first to correlate measures of fathers’ roles as providers to elevated T; however, our results complement and build on prior anthropological research that has shown that the well-documented differences in T between married fathers vs. single men and non-fathers do not necessarily extend to societies in which cultural models of partnering differ (e.g., polygyny is practiced) and roles for fathers involve little contact with young children (Gray, 2003; Muller et al., 2009).

To sum up, I began this chapter with a brief overview of human life history and current thinking regarding evolutionary-oriented theoretical frameworks (LHT, parental investment theory) that are commonly used to frame predictions regarding human paternal psychobiology. A fundamental idea underlying those frameworks is that if humans/hominins experienced selection favoring increased levels of paternal care, it would have required proximate (i.e., underlying physiological) mechanisms to help facilitate those shifts in male priorities and behavioral patterns. Based on cross-species comparisons, T is a likely candidate to function in that specific mechanistic role and is among the most widely studied hormones in research on human partnering/parenting and psychobiology. However, as I have reviewed above, there is increasing evidence from studies of T and, to a lesser extent, oxytocin (Abraham & Feldman, 2018; Gettler, Sarma, et al., 2019) that the patterning of human paternal biology potentially varies based on intersections between local contexts and systems of meaning (e.g., cultural norms), environmental conditions (i.e., ecology), and individual-level characteristics. It is important to note that these perspectives are not at odds with one another; rather, a number of lines of evidence are typically used by evolutionary-focused scholars to argue for plastic and facultative aspects of fathers’ involvement in our evolutionary past. Regardless of whether current variation in responses of fathers’ physiology to diverse parenting demands reflects selection in our evolutionary past or is merely an example of contemporary bio-behavioral phenotypic plasticity (Gettler, 2014), I would argue that to the extent that we think modeling between-father differences in these biological patterns is potentially important to understanding fathers’ effects on children and family systems, we need to continue to increase our attention to fathers’ own developmental experiences and individual dispositions as well as the prevailing cultural values of their communities and societies (Gettler, 2016).

Summary and Key Points

Humans have a relatively “slow” life history in many respects, especially the prolonged and costly periods of growth and development that are characteristic of our children. Yet, compared to closely related species, the way we raise our children is differentiated by the fact that in many families, parents and other caregivers commonly raise multiple dependent children simultaneously, rather than focusing on ensuring that each child reaches maturity before having another. Evolutionary and comparative perspectives on the emergence of this LH strategy help provide a theoretically grounded foundation for considering the important roles that human fathers have likely played throughout the history of our species and how selection may have shaped men’s psychobiology to help facilitate shifts toward committed partnering and parenting (Gettler, 2014; Gray & Anderson, 2010). In that vein, there are three key sets of points I would like to emphasize in closing this chapter.

First, evolutionary framing, including LHT, provides a critical foundation for proposing why men’s neurobiological and hormonal systems would have the functional capacity to respond to certain forms of partnering and parenting (Gettler, 2014). While the current literature on the biology of fatherhood is increasingly consistent with the idea that there is plasticity in fathers’ psychobiology based on social and ecological context, the neuroendocrine axis that produces T (for example) is evolutionarily ancient and highly conserved, being shared by all vertebrates, and is thus certainly not infinitely malleable (Gettler, 2016). Rather, evolutionary and comparative perspectives help us generate predictions about how hormones, such as T, may have been selected as mechanisms to mediate trade-offs between certain types of incongruent demands. In short, it is not theoretically appropriate to expect that fathers’ neuroendocrine systems will have the functional capacities to respond adaptively to all the demands of daily life. For example, a US-based father who is committed to being a sensitive and nurturing parent may also work in a business occupation in which he will achieve success by behaving highly competitively and hierarchically. These day-to-day demands are likely to be best optimized through different psychobiological profiles, which cannot necessarily be concomitantly expressed because our biology has constraints in its plasticity. Evolutionary perspectives are critical to understanding and modeling the plausible range of possibilities for parental physiology in contemporary environments and family systems, including in the face of political economic and structural constraints and challenges (Gettler, 2016).

Second, a fairly cohesive picture has emerged regarding men having lower T as fathers in the context of committed romantic partnerships and when they are more involved with direct caregiving for their children. However, these patterns are not ubiquitous across cultures, including in settings in which polygynous marriage is culturally sanctioned and practiced and those in which fathers’ roles generally do not include direct caregiving (Gettler, 2014; Gray, McHale, & Carre, 2017; van Anders, 2013). Moreover, recent research that my colleagues and I have published shows that in a society in which fathers’ roles as providers require risky behavior and represent pathways to achieving social status in a hierarchical community “better” fathers (by local definitions) have higher T than other fathers (Boyette et al., 2019; Gettler, Sarma, et al., 2019). Thus, there is much that remains poorly understood about the manifestation of human fathers’ biology (Abraham & Feldman, 2018; Gettler, 2014; Rosenbaum & Gettler, 2018), and researchers must be attentive to the importance of variation in cultural models of fatherhood and family life, including within subpopulations in large, diverse societies (Mazur, 2014), as this area of study continues to grow.

Finally, and on a related point, we need further insights on individual differences in men’s neurobiological and endocrine responsiveness from parenting and to understand the sources of that variation. Psychobiological frameworks increasingly emphasize the importance of early-life social-family environments in shaping the physiological pathways through which parenting is later expressed in adulthood (Bos, 2017; Gettler, 2016), while a small number of studies have begun exploring individual-level traits (e.g., personality dimensions; genetics) that predict differential outcomes of higher vs. lower T in the context of family life (Gettler et al., 2017; Sarma, Kuo, Bechayda, Kuzawa, & Gettler, 2018; van der Pol et al., 2019). For those of us who are interested in questions of parental biology and child outcomes, these more nuanced observations may be critical to helping solve the “conundrum” of how or if fathers’ psychobiological profiles are linked to the quality of their parenting and functioning within family systems.