developmental plasticity in fat patterning of ache children in response to variation in interbirth...

Original Research Article

Developmental Plasticity in Fat Patterning of Ache Children in Response toVariation in Interbirth Intervals: A Preliminary Test of the Roles of ExternalEnvironment and Maternal Reproductive Strategies

JACK BAKER,1* ANA MAGDALENA HURTADO,2 OSBJORN M. PEARSON,3 KIM R. HILL,2 TROY JONES,4

AND M. ANDERSON FREY3

1Bureau of Business and Economic Research, University of New Mexico, Albuquerque, New Mexico2School of Human Evolution and Social Change, Arizona State University, Tempe, Arizona3Department of Anthropology, University of New Mexico, Albuquerque, New Mexico4Department of Geography, University of New Mexico, Albuquerque, New Mexico

ABSTRACT A firm link between small size at birth and later more centralized fat patterning has been establishedin previous research. Relationships between shortened interbirth intervals and small size at birth suggest that mater-nal energetic prioritization may be an important, but unexplored determinant of offspring fat patterning. Potentialadaptive advantages to centralized fat storage (Baker et al., [2008]: In: Trevathan W, McKenna J, Smith EO, editors.Evolutionary Medicine and Health: New Perspectives. New York: Oxford) suggest that relationships with interbirthintervals may reflect adaptive responses to variation in patterns of maternal reproductive effort. Kuzawa ([2005]: Am JHum Biol 17:5–21; [2008]: In: Trevathan W, McKenna J, Smith EO, editors. Evolutionary Medicine and Health: NewPerspectives. New York: Oxford) has argued that maternal mediation of the energetic quality of the environment is anecessary component of developmental plasticity models invoking predictive adaptive responses (Gluckman and Han-son [2004]: Trends Endocrinol Metab 15:183–187). This study tested the general hypothesis that shortened interbirthintervals would predict more centralized fat patterning in offspring. If long-term maternally mediated signals are im-portant determinants of offspring responses, then we expected to observe a relationship between the average interbirthinterval of mothers and offspring adiposity, with no relationship with the preceding interval. Such a finding would sug-gest that maternal, endogenous resource allocation decisions are related to offspring physiology in a manner consistentwith Kuzawa’s description. We observed exactly such a relationship among the Ache of Paraguay, suggesting thatmaternally mediated in utero signals of postnatal environments may be important determinants of later physiology.The implications of these findings are reviewed in light of life history and developmental plasticity theories andour ability to generalize the results to other populations. Recommendations for further empirical research are brieflysummarized. Am. J. Hum. Biol. 21:77–83, 2009. ' 2008 Wiley-Liss, Inc.

A number of previous studies link in utero experience ofenergetic scarcity, typically measured as small size atbirth, to later body composition outcomes including morecentralized adiposity (Baker et al., 2008 for review). Ingeneral, biologists have been interested in these relation-ships in light of their potential linkage to the phenomenonof developmental plasticity (Gluckman and Hanson, 2004;Kuzawa, 2005), in which environmental factors experi-enced during development lead to life-long physiologicalalterations (Schlichting and Pigliucci, 1998). The presenceof developmental plasticity is ubiquitous throughout na-ture, and it is often cited as an adaptive mechanism fordealing with environmental heterogeneity (Jones, 2005;Levins, 1968; Schlichting and Piglucci, 1998; West-Eber-hard, 2003). Centralization of adiposity is a particularlynoteworthy phenomenon in this respect because it repre-sents a potential adaptation (Curio, 1973; Williams, 1966)to energetically marginal environments. Truncal fat cells(including both intra-abdominal and subcutaneous cells)are known to be more sensitive to lipolytic endocrine sig-nals than their counterparts in peripheral depots such asin the arms and legs (Arner, 1995; Belanger et al., 2006;Bjorntorp and Eden, 1996; Boivin et al., 2007; Evanset al., 1983; Hellmer et al., 1992; Kissebah et al., 1982;Rebuffe-Scrive et al., 1987, 1990; Rebuffe-Scrive, 1988;Wajchenberg, 2000). In an ancestral environment charac-terized by day-to-day fluctuation in energetic intakes that

are generally short term in nature and potentially modifi-able by additional foraging effort, enhanced centralizationof fat storage could improve energy availability for moder-ate-intensity activity in a fasting state, providing advan-tages in energy collection (Baker et al., 2008; Baker,2007).If this were the case, mechanisms whereby developing

fetuses detect signals of scarcity and respond with morecentralized fat storage could evolve as examples of thephenomenon of adaptive developmental plasticity (Bakeret al., 2008; Baker, 2007; Levins, 1968; Schlichting andPigliucci, 1998). Development of such a mechanism mightprovide additional benefits to standard selection of fixedgenetic responses, in that during times of plenty, a greateramount of resources could be invested in muscle tissueand associated physical performance, maximizing ener-getic collection through enhancing performance. The nec-essary signal might be provided by experiences of in utero

*Correspondence to: Jack Baker, Bureau of Business and EconomicResearch, 303 Girard Blvd. NE, MSC06 3510, 1 University of New Mexico,Albuquerque, NM 87106. E-mail: [email protected]

Received 21 July 2007; Revision received 18 July 2008; Accepted 21 July2008

DOI 10.1002/ajhb.20820

Published online 12 September 2008 in Wiley InterScience (www.interscience.wiley.com).

AMERICAN JOURNAL OF HUMAN BIOLOGY 21:77–83 (2009)

VVC 2008 Wiley-Liss, Inc.

scarcity (Baker, 2007; Baker et al., 2008; Kuzawa, 2005)which would be reflected in small size at birth; however,Kuzawa has rightly argued against the theoretical plausi-bility of a direct link between in utero experiences andlater physiology based on the idea that this direct experi-ence is an unreliable indicator of future environmentswithin a single lifetime. Instead, Kuzawa (2005, 2008) hasargued for the presence of ‘‘intergenerational inertia,’’ inwhich mothers provide mediated signals of environmentalquality to their offspring based on their own experiences.He provides a number of empirical results supporting theplausibility of this suggestion, some of the most compel-ling of which are observations that among Rhesus Maca-ques the mother’s own birth weight predicts that of heroffspring (Price and Coe, 2000) and that direct nutritionalsupplementation has been found to perform poorly inreducing low birth weight incidence in human populations(Ceesay et al., 1997).Shortened interbirth intervals have been repeatedly

linked to small size at birth (Conde-Aguelo et al., 2006;Khan et al., 1998; Zhu et al., 1999), suggesting that mater-nal reproductive decisions are related to offspring in uteroexperiences directly. If so, then measuring relationshipsbetween variation in interbirth intervals and offspring fatpatterning may provide a way to initially clarify if andhow mothers might mediate the environment of theirdeveloping offspring in a way that provides reliable sig-nals and allows the possible evolution of developmentallyplastic responses that are adaptive in nature. A poten-tially useful, but ultimately incomplete, starting point forthis assessment is the Smith and Fretwell (1974) model of

a trade-off between ‘‘quantity’’ and ‘‘quality’’ of offspring.The model provides a key, germaine point: an inherenttrade-off between increased reproduction and the fitnessof offspring exists because increasing reproductiondecreases the per-capita energetic availability to off-spring. The model was originally developed for under-standing litter and clutch size in which the per-capitadecrease in energy in relation to litter or clutch size is im-mediate and well linked to the literature on trade-offs inlife-history theory (Charnov and Ernst, 2006; Walkeret al., 2008). A simple prediction made in the spirit of thismodel, when extended to iteroparous, long-lived speciessuch as humans, would be that shortened interbirth inter-vals are directly related to energetic shortfalls in uteroexperienced by offspring and reflected in small size atbirth. The Smith-Fretwell model makes the assumptionthat the mother’s relative strategy of allocation to off-spring is time/situation invariant—she allocates resourcesevenly across all offspring. A simple extension thatrelaxes this assumption and allows mothers to weigh cur-rent and future reproduction against self-maintenanceand adjust endogenous resource allocation to these com-peting demands at each specific point in time (Ellison,2003; Stearns, 1992) makes a plausible link with what isknown about developmental plasticity. The overall frame-work of such a decision model is captured in Figure 1,which allows a disaggregation of the probability of concep-tion and the amount of resources allocated to each off-spring at each bout of reproduction. Such a disaggregationappears justified given that mothers clearly modulatethe probability of conception based on energetic status,

Fig. 1. Maternal reproductive investment decisions. At each time step, mothers must decide whether or not to reproduce, as well as howmuch to invest in self-maintenance versus reproduction.

78 J. BAKER ET AL.

American Journal of Human Biology

balance, and flux (Ellison, 2003) and it allows for mothersto adjust this from pregnancy-to-pregnancy based on long-term exposures as well as short-term environmentalinputs.

If we assume, then, that offspring will always respondto scarcity in utero by increasing centralization of fat pat-terning, then we may relate this outcome to differentmeasurements of interbirth intervals aimed at distin-guishing whether maternal reproductive strategy andassociated endogenous resource allocation is an importantdeterminant of offspring physiology. In doing so, we canmake a preliminary test of the idea that mothers maymediate energetic signals to their offspring, a necessarycomponent of models based on adaptive developmentalplasticity (Baker et al., 2008; Baker, 2007; Jones, 2005;Kuzawa, 2005, 2008). This will, perhaps, clarify whetheroffspring developmental responses are linked to short-term exogenously determined variation in resource avail-ability or to maternally mediated signals of scarcityreflecting her own priority rules for resource allocation(Baker et al., 2008; Kuzawa, 2005, 2008). In an attempt todiscriminate between exogenous environmental effectsand endogenous maternal resource allocation decisions,this study tests for relationships between two differentmeasures of interbirth intervals and offspring fat pattern-ing among Ache children 2 to 15 years of age. Interbirthintervals are operationalized as follows: (1) the interbirthinterval immediately preceding the birth of each studyparticipant or (2) the long-term ‘‘average’’ interbirth inter-val of the mother. Obviously, the two measures will berelated; however, if a relationship between increasinglycentralized fat patterning and the preceding interval isobserved in the absence of a relationship with the long-term average interval of the mother, it may suggest thatshort-term exogenously determined resource fluctuationsare the signal driving the developmentally plasticresponse. If a relationship with the long-term averageinterval is observed in the absence of a relationship withthe preceding interval, then it suggests that maternallymediated effects, perhaps similar to those envisioned inKuzawa’s intergenerational inertia model, are at work.The results of the study indicate initial support for theidea that maternal resource allocation preferences are im-portant in determining offspring fat patterning; however,they suggest a different relationship between these twovariables than was initially expected in the study. Thefindings suggest that specific strategies should beemployed in further research, which are reviewed. Impli-cations of the study for Kuzawa’s model, as well as forapplications of the Smith-Fretwell model to humans, arebriefly reviewed.

MATERIALS AND METHODS

Hypotheses and variables

We tested two hypotheses in this study. First, we testedthe prediction that shortened preceding interbirth inter-vals would result in more centralized fat patterning in off-spring. Second, we tested the prediction that shortened‘‘average’’ interbirth intervals of mothers would predictincreased centralization of fat patterning in offspring. Weoperationalized centralization of fat patterning using theratio of the subscapular skinfold measure to that of thetriceps (Malina et al., 1999). The subscapular skinfold, inparticular, has been repeatedly shown to be an adequate

measure of overall centralized fat patterning, capturingnot only subcutaneous fat storage, but also intra-abdomi-nal or ‘‘deep’’ fat storage through correlation (Ashwell,1985; Ferland et al., 1989; Fox et al., 2000; Ketel et al.,2007; Malina et al., 1999; Peters et al., 1994). In fact,Snijder et al. (2006), in reviewing various measures of cen-tralized adiposity, characterizes the ratio of the subscapu-lar to triceps skinfolds as the ‘‘best field measure of cen-tripetal fat patterning,’’ a point of view echoed elsewherein the literature (Ketel et al., 2007). Because of its knownrelationship with overall centralized fat patterning andmetabolic aberrations related to intra-abdominal obesity(Bjorntorp and Eden, 1996; Selby et al., 1989; Smoaket al., 1987; Yarnell et al., 2001), the subscapular/tricepsskinfold ratio continues to be used in clinical studies tocapture fat patterning variation (Monyeki et al., 2005;Moreno et al., 2007). As centralization of fat storageincreases, so does the subscapular/triceps skinfold ratio.This is reflected in Figure 2, where three ‘‘categories’’ ofcentralization of fat patterning are presented, with a caseof equivalent measurements yielding a ratio of 1, a casewith a ratio greater than 1 reflecting more centralized adi-posity, and a third case with a ratio less than 1 reflectingmore peripheral fat patterning. In our sample, the aver-age subscapular/triceps skinfold ratio was greater than 1(1.298). This ratio (and each individual component—thesubscapular and triceps skinfold measures) was related tothese two measures of interbirth intervals, with the expec-tation that the ratio would shift in the direction of a largervalue in response to shortened interbirth intervals.To adequately assess the predictions, we attempted to

control for a number of potential confounding effects.Mother’s weight (both the long-term average and theirweight immediately preceding the year of pregnancyresulting in the birth of the child) was included in anattempt to control for maternal nutritional status. Age atfirst birth was included as well to reflect quality of earlylife development of the mother, assuming that earlierreproduction would be linked to earlier menarche reflect-ing nutritional status during development. To control thepossibility of reduced energy availability to later-born off-spring of mothers—due to either maternal depletion(Tracer, 2003) or decreased per-capita energetic availability

Fig. 2. The ratio of the subscapular to the triceps skinfold willincrease as more fat is stored in the trunk. This may occur becausethe triceps skinfold is smaller, because the subscapular skinfold islarger, or more likely both.

79DEVELOPMENTAL PLASTICITY IN FAT PATTERNING


with increasing family size—we included both parity(completed fertility of the mothers at the time of thestudy) and birth order as a potential control variables aswell.In spite of these efforts, limitations in the data precluded

adequate control for a number of other potentially impor-tant confounding effects. First, genetic effects wereunmeasured because neither subscapular nor triceps skin-fold measures were available for either mothers or fathers.Genetic contributions to fat patterning are known to exist(Katzmarzyk and Bouchard, 1996) and are unmeasured inthe current study. A lingering problem related to maternalphysiology, first pointed out by Hill and Hurtado (1996)was an inability to control for variation in the intrinsicability of mothers to produce offspring. Given an equalamount of resources, some residual variation in this abilitybetween mothers could be present and remains unmeas-ured in this study. Additionally, a number of postnatalenvironmental factors related to energetic balance, includ-ing dietary saturated fat (Laitenen et al., 1992; Law et al.,1992) and activity levels were not measured in this study.Ideally, all of these variables would have been measuredwhile attempting to establish relationships between inter-birth intervals and offspring fat patterning.

Study population

The Ache, a group of recently settled hunter–gatherersresiding in rural Eastern Paraguay are ideal for studyingthe effects of maternally mediated signals of scarcity uponoffspring fat patterning. Previous research suggests thatthe Ache suffer universal chronic energetic stressorsacross development and throughout their lives (Hill andHurtado, 1996). Although it does not appear that Achechildren are born small (Hill, unpublished data), anunpublished analysis on the growth and nutrition of Achechildren suggested that as much as 2/3 of the populationbetween 2 and 15 years of age are stunted in staturegrowth (Baker, unpublished data). Unfortunately, lowresource availability and heavy work-loads characterizetheir lifestyle (Hill and Hurtado, 1996), suggesting that,research that might potentially be confounded by agreater access to resources could be successfully con-ducted there. An additional benefit for an evolutionaryand ecological study such as this is the fact that the Acheare historically hunter–gathers who only settled into sed-entary life within the past 35 years, and who still dependon foraging for a significant portion of their livelihood.This will allow us to assess these relationships within acontext more similar to the environment in which humansevolved, perhaps, allowing us to gain greater insight intothe nature of physiological adaptation in settings not asaffected by modern technology and healthcare as our own.

Data collection and analysis

We collected anthropometric measurements for 44 chil-dren in three Ache communities in Eastern Paraguayduring July and August, 2003. All measurements werecollected by a single data collector (JB) with significant ex-perience in anthropometry. Because of time constraintsintroduced by our need to provide medical care to a num-ber of seriously sick members of each community, meas-urements were taken on the right side only and repeatedmeasures and precision assessments were not performed.

Data was collected on pre-formed data collection sheets byan assistant data collector to avoid recollection bias in theshort period of time between measurement and recordingof the data. In each community, we attempted to gather auniversal sample of children aged 2 to 15 years. Thesedata were then combined with reproductive data from theAche Demographic Database, and with maternal anthro-pometric data from the Ache Anthropometric Database,both provided by Hill and Hurtado. The interbirth inter-vals (both short and long term) of the sample mothersappeared very similar to those of Ache mothers in general,averaging 31 months for the ‘‘average’’ interbirth intervaland 34 months for the preceding interval.These data were analyzed using standard regression

methodologies. This involved regressing the subscapular/triceps skinfold ratio, as well as the individual triceps andsubscapular skinfold measurements, against the immedi-ately preceding interbirth interval and the mother’s aver-age life-long interbirth interval at the time of observation.Controls for maternal condition (birth order, parity, andmother’s body weight) and other potential confounders(age and sex) were included in the analysis. Backwardselection was implemented in which an original modelwas fit including all predictor variables, noninformativepredictors were eliminated, one at a time, and the modelrefit (Neter et al., 2001). Although many epidemiologistsrecommend inclusion of all relevant ‘‘known’’ effects onthe outcome to be measured in a study (Aschengrau andSeage, 2003), there are potential drawbacks to such astrategy. The principal drawback to this approach is thatmany of these variables may be noninformative (in thestatistical sense, they do not improve model fit, diagnosticmeasures, or the ‘‘r-squared’’ value) in the particularstudy setting utilized in this study. If so, these noninfor-mative predictors can bias the regression parameters(Greene, 2003) indicating a sacrifice would be made in thestudy’s ability to accurately assess the relationship to bemeasured. Because a power calculation suggested onlylimited power to detect real effects of the independentvariable (b 5 0.21937), we decided to eliminate these non-informative variables in this preliminary study in favor ofmaximizing our ability to accurately estimate the parame-ters in the regression model. Thus, final models reportedhere included only those variables shown to have an in-formative value. The final models were assessed graphi-cally for multicolinearity in the predictors (using correla-tion plots between the predictors) as well as formallyusing Variance Inflation Factors (Neter et al., 2001). Over-all model bias introduced by left-out variable error(Mauro, 1990) was assessed using the Mallow’s C-p statis-tic. Overall diagnostics of model fit include residual by fitplots and normal plots of the residuals (Neter et al., 2001).All diagnostics suggested appropriate fit of the final mod-els, with little indication of left-out variable error.

RESULTS

The analyses suggested no relationship between thesubscapular/triceps skinfold ratio with the interval pre-ceding the birth of the study subject (F 5 0.32, P 50.82935). The analysis suggested significant relationshipsbetween the mother’s average interbirth interval and boththe subscapular/triceps skinfold ratio (b 5 0.14047, t 52.68727, P 5 0.0109) and the triceps skinfold itself (b 522.34349, t5 22.18567, P5 0.0356); however, the direction

80 J. BAKER ET AL.


of the observed effect was in the opposite direction fromthat originally predicted to characterize a relationshipbetween the interbirth interval and a signal of scarcity tothe offspring. No effect of interbirth intervals (either pre-ceding or average) was observed in conjunction with thesubscapular skinfold (F 5 0.62, P 5 0.6095). As the aver-age interbirth intervals of Ache mothers increased, so didthe subscapular/triceps skinfold ratio (Table 1, Fig. 3).This was due to a reduction in the triceps skinfold as themother’s average interbirth interval increased (Table 2,Fig. 4). The final model relating the average interbirthinterval to the subscapular/triceps skinfold ratioaccounted for 39.78% of the variance in the ratio (F 59.37, P 5 0.0001). The final model relating average inter-birth intervals to the triceps skinfold accounted for31.26% (F 5 6.76, P 5 0.001) of the variation in the tricepsskinfold.

The model predicting the subscapular/triceps skinfoldratio also included mother’s weight and the offspring’sbirth order. As birth-order increased, so did the ratio (b 50.132971, t 5 4.41, P 5 0.001), perhaps, suggesting a rela-tionship between offspring centralized adiposity andpotential maternal depletion. As mother’s average weightincreased, the subscapular/triceps skinfold ratio actuallydecreased (b 5 20.017424, t 5 23.05, P 5 0.0043), sug-gesting that increased mother’s average weight, perhaps,reflecting either increased nutritional status, baselinephysiological capacities, early-life environment, or allthree could be important in understanding mother/off-spring relationships as well. No interaction effect wasobserved, however, suggesting a separate role for thisfrom that suggested for the interbirth interval itself. The

model predicting reduced triceps skinfold as the drivingforce in the increase of the ratio included the same twovariables, birth order (b 5 22.17703, t 5 23.51897, P 50.0012), and mother’s average weight (b 5 0.34208, t 52.91923, P 5 0.0061). Figures 2 and 3 graphically repre-sent the two results: increasing subscapular/triceps skin-fold and a decreasing triceps skinfold as average inter-birth intervals increase.

DISCUSSION

At first blush, the results of the study seem to indicatethat neither the preceding nor the average interbirthintervals are related to offspring adiposity in the wayenvisioned by this study. We think, however, that whenproperly evaluated the results suggest another conclusion.If we assume that offspring consistently respond to inutero energetic scarcity by becoming more centrallyadipose, then the observed results suggest that maternalendogenous rules for prioritizing the allocation of re-sources—presumably reflected in the average interbirthinterval—are directly related to offspring developmentalplasticity in the form of increasingly centralized adiposity.The reflection of this in the subscapular/triceps skinfoldratio is due to decreases in the triceps skinfold, suggestinggreater emphasis upon storage of fat in truncal cells thatare more sensitive to lipolytic endocrine signals. Althoughwe anticipated that shortened interbirth intervals wouldbe related to increasing central adiposity in the offspring,the observation that extended interbirth intervals arerelated to centralized fatness, leads directly to the conclu-sion that maternal endogenous prioritization is poten-tially responsible for signaling offspring of environmentalscarcity. This may be true because of the nature of a very

Fig. 3. As the mother’s average interbirth interval increases, sodoes the subscapular/triceps skinfold ratio in her offspring. [Colorfigure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

TABLE 2. Results of multiple regression: triceps skinfold (millimeters)by mother’s average interbirth interval (years)

Parameter Estimate Standard error t P-value

Constant 4.43327 7.78746 0.569283 0.5728Average IBI 22.34349 1.07221 22.18567 0.0356Birth order 22.17703 0.618655 23.51897 0.0012Mother kg 0.34208 0.117182 2.91923 0.0061

F 5 6.76 (P5 0.001), R-squared5 31.26 %.

Fig. 4. As the mother’s average interbirth interval increases, thetriceps skinfold decreases. [Color figure can be viewed in the onlineissue, which is available at www.interscience.wiley.com.]

TABLE 1. Results of multiple regression: subscapular/triceps skinfoldratio by mother’s average interbirth interval (years)

Parameter Estimate Standard error t P-value

Constant 1.48826 0.379484 3.92181 0.0004Average IBI 0.140407 0.0522487 2.68727 0.0109Birth order 0.132971 0.0301471 4.41073 0.0001Mother kg 20.017424 0.00571027 23.05134 0.0043

F 5 9.37 (P5 0.0001), R-squared5 39.78 %.



general decision model underlying maternal reproductivedecisions (see Fig. 1). At each step beyond the birth of achild, the mother decides whether to reproduce or not andthis clearly reflects her physiological preferences for cur-rent/future reproduction (step 1 of the model in Fig. 1).Once the decision to reproduce has been made, however,she will also decide to what degree she allocates resourceseither to self-maintenance or to the developing fetus.(Step 2 in the model). Given this set of relationships,mothers with more conservative preferences (those whotend to prefer self-maintenance in general) may be likelyto reproduce less frequently as well as to allocate fewerrelative resources to gestating offspring. If so, then lessfrequent reproduction could clearly be linked to signals ofscarcity to offspring, resulting in developmental plasticityand observed increases in centralization of adiposity. It isalso true, however, that mothers may choose to reproduceless frequently but to allocate an equal amount of re-sources to each child. If so, then we might expect to see ei-ther no evidence of birth order effects or a dampenedeffect. In this study, we observed increasing centralizedadiposity with birth order, driven by decreases in the tri-ceps skinfold, that was both statistically significant andphysiologically important. This observation suggests thatmaternal resource constraints do exist, but only furthertests will decide whether this relationship is due to mater-nal depletion, or relative devaluing of reproductive effortlater during reproductive life.Although intriguing, these results are preliminary in

nature. The inability within this study to control for anumber of confounding effects limits the inferences thatmay be drawn from it. Further cross-sectional studies ofthis kind will only go so far in discerning the relativeeffects of exogenous environment and endogenous/mater-nal factors on shaping offspring body composition. Onlyfurther studies that are longitudinal in nature and pos-sess the ability to control for maternal nutrition duringpregnancy, variation in maternal physiological capacitiesand her own associated development, and a variety ofpostnatal environmental influences on adiposity willclearly determine whether models invoking maternalmediation of environmental cues (such as Kuzawa’s inter-generational inertia model) are adequate explanations forlinks between in utero experiences and later offspring adi-posity. Ideally, such a study would be designed with evolu-tionary logic in mind (Baker et al., 2008), yielding specific,testable predictions that advance evolutionary theorizingabout life-history and physiological variation in humans.The results of this research also suggest that the Smith-Fretwell model may be a useful framework for under-standing maternal/child health and the determinants ofdevelopmental plasticity in offspring physiology, but notwithout modification. Relaxing assumptions about thetemporally static nature of maternal decisions to repro-duce or not, or how much of her available energeticresources to invest in either reproduction or self-mainte-nance proved to be useful in the current study. Humansare a long-lived, iteroparous species whose decisionsshould be expected to be dynamic, varying over time basedon current as well as anticipated conditions (Charles-worth, 1980; Levins, 1968). At the same time, they arealso likely to be subjected to constraints imposed by theprocess of canalization, in which phenotypes become pro-gressively less flexible up to adulthood (Schmalhausen,1949; Shlichting and Pigliucci, 1998). Ideally, a clearer

understanding of how the exogenous environment, mater-nal reproductive decisions and resource prioritization,and offspring fat patterning relate will require extensionsof both the Smith-Fretwell model and developmental plas-ticity theory to elucidate the role of how mother’s resourceallocation preferences are shaped by her own developmen-tal experiences.

ACKNOWLEDGMENTS

An earlier version of this manuscript was greatlyenhanced by the feedback of two anonymous reviewers aswell as Peter Ellison. Although we want to acknowledgethese contributions, any errors or omissions are of ourown.

LITERATURE CITED

Arner P. Differences in lipolysis between human subcutaneous and omen-tal adipose tissues. 1995. Ann Med 27:435–438.

Aschengrau A, Seage G. 2003. Essentials of epidemiology in public health,2nd ed. New York: Barlett and Jones.

Ashwell M, Cole TJ, Dixon AK. 1985. Obesity: new insight into the anthro-pometric classification of fat distribution shown by computed tomogra-phy. BMJ 290:1692–1694.

Baker J. 2007. The evolutionary ecology of thrifty metabolism: early-lifesignals of environmental instability and later body composition inhumans. Ph.D. Dissertation Manuscript, University of New Mexico.

Baker J, Hurtado AM, Pearson OM, Jones T. 2008. Evolutionary medicineand obesity: developmental adaptive responses in human body composi-tion. In: Trevathan W, McKenna J, Smith EO, editors. Evolutionarymedicine and health: New Perspectives. New York: Oxford.

Belanger C, Hould FS, Lebel S, Biron S, Brochu G, Tchernof A. 2006.Omental and subcutaneous adipose tissue steroid levels in obese men.Steroids 71:674–682.

Bjorntorp P, Eden S. 1996. Hormonal influences on body composition. In:Roche AF, Lohman SB, editors. Human body composition. Champaign:Human Kinetics. p 424–440.

Boivin A, Brochu G, Marceau S, Marceau P, Hould FS, Tchernof A. 2007.Regional differences in adipose tissue metabolism in obese men. Metabo-lism 56:533–540.

Ceesay SM, Prentice AM, Cole TJ, Foord F, Weaver LT, Poskitt EM,Whitehead RG. 1997. Effects on birth weight and perinatal mortality ofmaternal dietary supplements in rural Gambia: 5 year randomised con-trol trial. BMJ 315:786–790.

Charlesworth B. 1980. Evolution in age-structured populations. Oxford:Oxford UP.

Charnov E, Ernst SKM. 2006. The offspring-size/clutch-size trade-off inmammals. Am Nat 167:278–582.

Conde-Agudelo A, Rosas-Bermudez A, Kafury-Goeta AC. 2006. Birth spac-ing and the risk of adverse perinatal outcomes: a meta-analysis. JAMA295:1809–1823.

Curio E. 1973. Towards a methodology of teleonomy. Experientia 29:1045–1059.

Ellison P. 2003. Energetics and reproductive effort. Am J Hum Biol15:342–351.

Evans D, Hoffman R, Kalkhoff R, Kissebah A. 1983. Relationship of andro-genic activity to body fat topography, fat cell morphology, and metabolicaberrations in premenopausal women. J Clin Endocrinol Metab 57:304–310.

Ferland M, Despres JP, Tremblay A, Pinault S, Nadeau A, Moorjani S,Lupien P, Theriault G, Bouchard C. 1989. Assessment of adipose tissuedistribution by computed axial tomogaphy in obese women: associationwith body density and anthropometric measurements. Br J Nutr61:139–148.

Fox KR, Peters DM, Sharpe P, Bell M. 2000. Assessment of abdominal fatdevelopment in young adolescents using magnetic resonance imaging.Int J Obes 24:1653–1659.

Greene WH. 2003. Econometric analysis, 5th ed. Upper Saddle River:Prentice Hall.

Gluckman PD, Hanson MA. 2004. The developmental origins of the meta-bolic syndrome. Trends Endocrinol Metab 15:183–187.

Hellmer J, Marcus C, Sonnenfeld T, Arner P. 1992. Mechanisms for differ-ences in lipolysis between human subcutaneous and omental fat cells.J Clin Endocrinol Metab 75:15–20.

Hill K, Hurtado AM. 1996. Ache life history: the ecology and demographyof a foraging people. New York: Aldine.

82 J. BAKER ET AL.


Jones JH. 2005. Fetal programming: adaptive life-history tactics or mak-ing the best of a bad start? Am J Hum Biol 17:22–33.

Katzmarzyk P, Bouchard C. 1996. Genetic influences on human body com-position. In: Roche AF, Lohman SB, editors. Human body composition.Champaign: Human Kinetics.

Ketel IJG, Volman MNM, Seidell JC, Stehouwer CDA, Twisk JW, LambalkCB. 2007. Clinical study: Superiority of skinfold measurements andwaist over waist-to-hip ratio for determination of body fat distribution ina population-based cohort of caucausian dutch adults. Eur J Endocrinol156:655–661.

Khan KS, Chien PF, Khan NB. 1998. Nutritional stress and reproduction:A cohort study over two consecutive pregnancies. Acta Obstet GynecolScand 77:395–401.

Kissebah A, Vydelingum N, Murray R, Evans D, Hartz A, Kalkhoff R,Adams P. 1982. Relation of body fat distribution to metabolic complica-tions of obesity. J Clin Endocrinol Metab 54:254–260.

Kuzawa CW. 2005. Fetal origins of developmental plasticity: are fetal cuesreliable predictors of future nutritional environments? Am J Hum Biol17:5–21.

Kuzawa CW. 2008. The developmental origins of adult health: intergenera-tional inertia in adaptation and disease. In: Trevathan W, McKenna J,Smith EO, editors. Evolutionary medicine and health: New Perspec-tives. New York: Oxford.

Laitenen J, Pietilainen K, Wadsworth M, Sovio U, Jarvelin MR. 2004. Pre-dictors of abdominal obesity among 31-year-old men and women born innorthern Finland in 1966. Eur J Clin Nutr 58:180–190.

Law CM, Barker DJ, Osmond C, Fall CH, Simmonds SJ. 1992. Earlygrowth and abdominal fatness in adult life. J Epidemiol CommunHealth 46:184–186.

Levins R. 1968. Evolution in changing environments: some theoreticalexplorations. Princeton: Princeton University Press.

Malina RM, Koziel S, Bielicki T. 1999. Variation in subcutaneous adiposetissue distribution associated with age, sex, and maturation. Am J HumBiol 11:189–200.

Mauro R. 1990. Understanding L.O.V.E. (left out variables error): amethod for estimating the effects of ommitted variables. Psychol Bull108:314–329.

Monyeki KD, Kemper HCG, Makgae PJ. 2005. The association of fat-pat-terning with blood pressure in rural South African children: The EllisrasLongitudinal Growth and Health Study. Int J Epidemiol 35:114–120.

Moreno LA, Mesana MI, Gonzalez-Gross M, Gil CM, Ortega FB, Fleta J,Warnberg J, Leon JF, Marcos A, Bueno M, the AVENA Study Group.2007. Body fat distribution reference standards in Spanish adolescents:The AVENA study. Int J Obes 31:1798–1805.

Neter J, Kutner MH, Nachtsheim CJ, Wasserman W. 2001. Applied linearstatistical models. Boston: McGraw-Hill.

Peters DM, Fox KR, Armstrong N, Sharpe P, Bell M. 1994. Estimation ofbody fat and body fat distribution in 11-year-old children using magnetic

resonance imaging and hydrostatic weighing, skinfolds and anthropom-etry. Am J Hum Biol 6:237–243.

Price KC, Coe CL. 2000. Maternal constraint on fetal growth patterns inthe rhesus monkey (Macaca mulatta): the intergenerational linkbetween mothers and daughters. Hum Reprod 15:452–457.

Rebuffe-Scrive M. 1988. Steroid hormones and distribution of adipose tis-sue. Acta Med Scand 723:143–146.

Rebuffe-Scrive M, Bronnegard M, Nilsson A, Eldh J, Gustafson JA, Bjorn-torp P. 1990. Steroid hormone receptors in human adipose tissues. J ClinEndocrinol Metab 71:1215–1219.

Rebuffe-Scrive M, Lonnroth P, Marin P, Wesslau C, Bjorntorp P, Smith U.1987. Regional adipose tissue metabolism in men and postmenopausalwomen. Int J Obes 11:347–355.

Schlichting CD, Pigliucci M. 1998. Phenotypic evolution: a reaction normperspective. New York: Sinauer.

Schmalhausen II. 1949. Factors of evolution. Philadelphia: Blackstone.Selby JV, Friedman G, Queensberry CP. 1989. Precursors of essential

hypertension: the role of body fat distribution pattern. Am J Epidemiol129:43–53.

Smith CC, Fretwell SD. 1974. The optimal balance between size and num-ber of offspring. Am Nat 108:499–506.

Smoak C, Burke G, Webber L, Harsha DW, Srinivasan S, Berenson G.1987. Relation of obesity to clustering of cardiovascular disease risk fac-tors in children and young adults: The Bogalusa Heart Study. Am J Epi-demiol 125:364–372.

Snijder MB, van Dam RM, Visser M, Seidell JC. 2006. What aspects ofbody fat are particularly hazardous and how do we measure them? Int JEpidemiol 35:83–92.

Stearns SC. 1992. The evolution of life histories. New York: Oxford Univer-sity Press.

Tracer D. 2002. Somatic versus reproductive energy allocation in PapuaNew Guinea: life history theory and public health policy. Am J Hum Biol14:621–626.

Wajchenberg B. 2000. Subcutaneous and visceral adipose tissue: their rela-tion to the metabolic syndrome. Endocr Rev 21:697–738.

Walker RS, Gurven M, Burger O, Hamilton MJ. 2008. The trade-offbetween number and size of offspring in humans and other primates.Proc Biol Sci 275:827–833.

West-Eberhard MJ. 2003. Developmental plasticity and evolution. NewYork: Oxford University Press.

Williams GC. 1966. Adaptation and natural selection. New York: Sinauer.Yarnell JWG, Patterson CC, Thomas HF, Sweetnam PN. 2001. Central

obesity: predictive value of skinfold measurements for subsequentischaemic heart disease at 14 years follow-up in the Caerphilly Study.Int J Obesity 25:1546–1549.

Zhu BP, Rolfs RT, Nangle BE, Horan JM. 1999. Effect of the intervalbetween pregnancies on perinatal outcomes. N Engl J Med 340:589–594.



developmental plasticity in fat patterning of ache children in response to variation in interbirth...

Documents