Are parental condition transfer effects more widespread than is currently appreciated?

Much research has demonstrated that the environment experienced by organisms during development can induce phenotypic changes that persist for multiple generations (Bashey, 2006; Jablonka and Raz, 2009; Bonduriansky et al., 2012; Salinas and Munch, 2012; Walsh et al., 2015, 2016; Bell and Hellmann, 2019; Donelan et al., 2020). This transgenerational plasticity occurs when the environment experienced by parents induces phenotypic changes in offspring and possibly future generations. One of the most prominent examples of a transgenerational response are ‘maternal effects’ or, more broadly, ‘parental effects.’ Parental effects manifest anytime the parental environment influences offspring phenotypes via impacts on offspring provisioning, hormones, behavior or epigenetic effects etc. (Bernardo, 1996; Grafen, 1988; Mousseau and Fox, 1998; Gustafsson et al., 2005; Marshall and Uller, 2007; Taborsky et al., 2007; Leips et al., 2009; Burton et al., 2020). A substantial amount of research has focused on the maternal role in parental effects and explored the extent to which mothers respond to predictable shifts in environmental conditions by modifying offspring phenotypes to best match future conditions (Uller et al., 2013). While there are clear examples of such ‘anticipatory maternal effects’ (Fig. 1A) (Agrawal et al., 1999; Galloway and Etterson, 2007; Salinas and Munch, 2012; Lind et al., 2020), the generality of this maternal effect remains unclear (Uller et al., 2013). This is because many organisms experience environmental conditions that theory predicts should favor the evolution of anticipatory effects (Marshall and Uller, 2007; Ezard et al., 2014; Kuijper and Hoyle, 2015; Leimar and McNamara, 2015; Uller et al., 2015) but a meta-analysis described the empirical evidence for anticipatory effects as ‘weak’ (Uller et al., 2013). What explains this disconnect between theory and the empirical literature?

‘Condition transfer effects’ represent an alternative class of parental effects that are less well studied than anticipatory effects (Bonduriansky and Crean, 2018). These effects occur when parents that experience high-quality conditions as they develop produce ‘fitter’ offspring across all environments that the offspring may encounter. Here, maternal and/or parental condition is ‘transferred’ to offspring such that there is a positive relationship between parental condition and offspring performance (Fig. 1B) (Bonduriansky and Crean, 2018; Bonduriansky, 2021). Silver spoon effects, where mothers that develop in high-quality habitats always produce higher-quality offspring, represent another term for a condition transfer effect (Grafen, 1988). Condition transfer effects represent a class of maternal effects that do not rely upon external environmental signals but are instead a byproduct of past environmental quality (Grafen, 1988; Van de Pol et al., 2006). Predictable shifts in environmental conditions will often covary with shifts in resources (or environmental quality etc.). Life history theory has long predicted that variation in resources is central to the evolution of maternal provisioning strategies (Charlesworth, 1980; Stearns and Koella, 1986; Kozlowski and Uchmanski, 1987; Kozlowski and Wiegert, 1987). Thus, condition transfer effects may provide an answer to the paradoxical lack of evidence for anticipatory maternal effects.

In this Review, we provide an overview of the current understanding of condition transfer effects in the literature. We specifically review studies that manipulate naturally occurring environmental stressors in the parental generation and then evaluate the fitness of offspring in response to exposure to that same stressor. We address three main questions: (1) Are condition transfer effects more widespread than is currently appreciated? (2) Which environmental stressors are most commonly associated with condition transfer effects? (3) Are there factors that influence the induction of condition transfer effects?

Early interest in environmentally induced parental effects focused on the relationship between mothers and offspring (Bell and Hellmann, 2019). One form of a maternal effect is the connection between environmental quality, maternal provisioning of offspring, and offspring performance. Given that resources are typically limited for reproduction, it was recognized early on that selection should favor a balance between the production of many, small offspring versus the production of a few, large individuals (Lack, 1954; Smith and Fretwell, 1974). This led to the realization that spatial or temporal variation in environmental conditions may exert selection on patterns of maternal provisioning of offspring. In general, selection is expected to maximize maternal fitness as opposed to offspring fitness, especially in species that lack parental care and that reproduce many times (Marshall and Uller, 2007). As a result, there are several ways in which environmental conditions can modify the link between female and/or parental development and the traits of offspring. This includes the aforementioned anticipatory and condition transfer effects but also ‘selfish’ maternal effects that maximize maternal fitness at the expense of offspring fitness (Scheirs et al., 2000; Mayhew, 2001) and bet hedging (Shama, 2015; Tufto, 2015; Furness et al., 2015). Below, we provide a more thorough introduction into anticipatory effects and condition transfer effects as they are the focus of this Review.

A substantial body of research has explored the extent to which maternal responses to environmental conditions enhance the fitness of offspring (reviewed by Marshall and Uller, 2007; Uller et al., 2013). These maternal effects are based upon the assumption that mothers ‘anticipate’ the environmental conditions that offspring will experience (i.e. high quality versus low quality, harsh versus benign, stressful versus non-stressful) and that they will then modify the phenotype of offspring accordingly (Marshall and Uller, 2007; Allen et al., 2008; Burgess and Marshall, 2014). Selection is expected to favor anticipatory maternal effects when: (1) environmental conditions shift predictably between generations, (2) cues reliably predict offspring conditions, and (3) the cost of maternal cue induction and transmission is low (Marshall and Uller, 2007). The empirical approach to studying anticipatory maternal effects typically utilizes factorial manipulations of maternal and offspring conditions. One example of a result supporting the existence of anticipatory maternal effects is evident by a significant maternal×offspring interaction such that offspring perform ‘best’ when reared in the same conditions that their mother experienced (Fig. 1A) (Dey et al., 2016). An additional possibility is that parents adjust the phenotype of offspring to an alternative environment when conditions change predictably between generations. The key result that supports the existence of an anticipatory response is that mothers that experience a low-quality or stressful environment will produce offspring with higher fitness than the offspring of mothers from high-quality environments when the offspring are exposed to the low-quality conditions (Fig. 1A). Some studies have provided clear evidence for the existence of anticipatory effects (Agrawal et al., 1999; Galloway and Etterson, 2007; Salinas and Munch, 2012; Lind et al., 2020).

Condition transfer effects are maternal effects that are based upon the notion that an organism's condition reflects its quantity of energy reserves (Andersson, 1982; Rowe and Houle, 1996). One assumption is that ‘high-condition’ individuals have increased energy reserves, which, in turn, leads to the production of offspring with higher fitness. These maternal condition transfer effects then occur when investment in offspring quality is positively correlated with parental condition (Marshall and Uller, 2007; Bonduriansky and Crean, 2018). Here, high-condition parents ‘transfer’ their condition to offspring, and offspring from high-condition parents outperform offspring from low-condition parents regardless of the environment that offspring experience. This includes offspring of high-condition mothers having higher fitness especially in low-quality or stressful environments (Fig. 1B). Note that it may be possible to have a condition transfer effect and anticipatory effect operating at the same time if the effects of the environment on parents change over time (see Monaghan, 2008).

Condition transfer effects have historically received less attention than anticipatory effects (Bonduriansky and Crean, 2018; Bonduriansky, 2021). This is, in part, because researchers have considered condition transfer effects to represent a non-adaptive byproduct of parental quality. Yet, Bonduriansky and Crean (2018) made the important point that any form of parental effects will be adaptive if they increase parental (lifetime) reproductive success (see also Marshall and Uller, 2007). As a result, condition transfer effects have the potential to be adaptive because parents that experience high-quality conditions during development will likely produce offspring with higher fitness and/or produce more offspring throughout their lifetime (see Parker, 2002; Coulson and Porter, 1985; Ciconia, 2001; Van de Pol et al., 2006). In addition to being adaptive, condition transfer effects have the potential to be more widespread than anticipatory maternal effects. One reason is because anticipatory effects require the evolution of mechanisms (i.e. neural, sensory) that allow organisms to accurately receive and process environmental signals and modify the traits of offspring (Bonduriansky and Crean, 2018; Bonduriansky, 2021). In contrast, the expression and evolution of condition transfer effects more simply require a link between parental condition and investment in eggs/offsping and/or the transmission of factors that influence offspring performance (i.e. hormones, epigenetic effects). As a result, condition transfer effects are potentially more pervasive than is currently appreciated.

The main goal of this Review was to provide new insight into condition transfer effects. It was not intended to be comprehensive but instead highlight whether (or not) condition transfer effects are prevalent and worthy of further study. This Review also represents an extension of the earlier review on anticipatory maternal effects by Uller et al. (2013). This is because studies that manipulate environmental conditions during parent and offspring generations and thereby test for anticipatory effects can also test for condition transfer effects (see below). As a result, we first collated all of the papers that were included in Uller et al. (2013). We then used the same keyword search in ISI Web of knowledge as used by Uller et al. (2013). This search included the following terms: (maternal effect* OR parental effect* OR transgenera*plastic*) AND (experiment*) AND (environment*) AND (offspring OR fitness). This search was conducted in March 2023 and spanned the years 2011–2023. The key difference between anticipatory versus condition transfer effects relates to the fitness of offspring from parents exposed to stressful or low-quality conditions (Fig. 1). As a result, there were several criteria that we used when determining whether to include a paper in this Review. The criteria were: (1) the study reared the parental and offspring generation on contrasting environmental conditions (or a range of conditions). (2) We were able to classify one of the environments as low quality or stressful. This is important because studies that imposed stressful treatments would be more likely to generate parents of contrasting condition. This would, in turn, allow us to better distinguish anticipatory versus condition transfer effects. This eliminated studies that, for instance, manipulated temperature but did not specify whether the increase or decrease in temperature was likely to be stressful. (3) The study assessed the fitness of offspring in the low-quality or stressful environment. This allowed us to determine if high-condition parents transfer their condition to offspring in low-quality conditions. The lone exception is that we did include two studies that exposed a parental generation to predator cues but did not manipulate the offspring generation. (4) The study manipulated a naturally occurring environmental stressor. This eliminated studies that involved, for instance, hormone manipulations.

Based upon these search parameters and criteria, this Review included 55 papers. We defined specific criteria for results supporting anticipatory versus condition transfer effects. We considered a paper as providing evidence for anticipatory effects when the offspring of mothers or parents from the low-quality or stressful environment had higher fitness in the low-quality environment when compared with fitness of offspring from mothers/parents from the high-quality environment (Fig. 1A). The opposite trend (mothers or parents from high-quality environments producing more fit offspring) instead provided evidence for condition transfer effects (Fig. 1B). It is also possible for studies to provide evidence for both anticipatory and condition transfer effects. This is, in part, because it is difficult to design experiments that will unambiguously tease apart anticipatory effects from condition transfer effects as both could be operating simultaneously (Enqqvist and Reinhold, 2016). Studies could provide evidence for both forms of parental effects if one trait provides evidence for an anticipatory effect while another trait is consistent with the criteria for a condition transfer effect. Also note that the expression of parental effects can be influenced by contributions from mothers and fathers. As a result, we also tracked whether studies exposed one or both parents and also whether studies were able to disentangle parental effects due to mothers versus fathers (see Table 1).

There is widespread evidence for condition transfer effects (Table 1). Greater than 50% (29/55 studies) of the empirical studies included in this review observed evidence for condition transfer effects (Table 1). In such studies, there was at least one example where the trait value of the offspring of parents that experienced high-quality conditions exceeded the trait values of offspring of parents that experienced low-quality conditions when the offspring were in the stressful environment (Fig. 1B). That is, the fitness of offspring of high-condition parents was higher than the fitness of offspring of low-condition parents in the low-quality or stressful environment. The evidence for condition-dependent transfer effects is taxonomically diverse as there are examples in plants, aquatic invertebrates, terrestrial invertebrates, fish and reptiles (Table 1). Furthermore, we identified 35 studies that manipulated resource quality or quantity (Table 1). This included studies that directly manipulated resources (or water for plants) or proxies for resources such as density. Condition transfer effects were more commonly observed (21/35 or 60%) than anticipatory effects (16/35=46%) in such studies. Some studies that manipulated resources revealed very clear cases of condition transfer effects. For example, Bonduriansky and Head (2007) manipulated larval diet quality in the parent and offspring generations in the fly (Telostylinus angusticollis). They showed that high-condition mothers produced offspring that developed faster on the low-quality food treatment when compared with the offspring of low-condition mothers. Similarly, Beyer and Hambright (2017) reared multiple generations of rotifers (Brachionus calyciflorus) on high-quality (Chlamydomonas) versus low-quality (Microcystis) algae and found that the offspring of mothers fed high-quality food had higher fitness (intrinsic rate of increase) on the low-quality food when compared with the offspring of mothers fed Microcystis.

The studies that document condition transfer effects are not limited to manipulation of resources (Table 1). Of the 20 studies that manipulated environmental factors other than resources, 7/20 (35%) detected evidence for condition transfer effects. For example, Fernández-González et al. (2011) exposed water fleas (Daphnia magna) to sub-lethal concentrations of copper. The offspring of parental Daphnia that were not exposed to copper were larger and produced more offspring compared with the offspring of parents that were exposed to copper. Such differences were only apparent when the offspring were exposed to copper. In a separate study on Daphnia magna, Mikulski and Pijanowska (2017) found that mothers that were exposed to fish predator kairomones produced offspring that were smaller at maturation and exhibited lower reproductive investment when compared with offspring from control mothers (when the offspring generation was reared in the presence of predator cues).

It was not uncommon for studies to document the presence of condition transfer effects for one trait but an anticipatory response for another trait (Table 1). Wang et al. (2017) manipulated resource availability in the lizard (Eremias multiocellata) and observed that the offspring of high-condition mothers (i.e. fed high food levels) exhibited a faster sprint speed on low food compared with the offspring of low-condition mothers, clearly showing a condition transfer effect. Conversely, the offspring of low-condition mothers grew faster on low food when compared with the growth rates of high-condition mothers, indicating an anticipatory effect. Sultan et al. (2009) observed contrasting responses to drought stress in two closely related species of plants. In one species (Polygonum persicaria), drought stressed parents produced offspring with longer roots and larger seedling biomass in dry soil versus plants from parents grown in moist soil. Such results support an anticipatory response. In Polygonum hydropiper, the offspring of drought stressed parents had shorter roots and were smaller in dry soil, which is consistent with a condition transfer effect (non-stressed parent produced offspring with higher fitness). Given that P. persicaria is found in moist and dry sites whereas P. hydropiper is limited to moist sites, these results also indicate that contrasting ecological conditions may favor distinct forms of maternal effects.

The evidence for condition transfer effects was, of course, not universal. Several studies did not find any evidence for condition transfer effects or anticipatory maternal effects (Table 1). In some cases, this does not mean that these studies did not reveal any maternal effects but instead did not find evidence for anticipatory or condition transfer effects as defined in the review criteria for this study (see Coslovsky and Richner, 2012; Coakley et al., 2018; Botterill-James et al., 2019; Mojzes et al., 2021). For instance, Botterill-James et al. (2019) manipulated resource availability during gestation and found that low-condition mothers (i.e. fed low food) produced offspring that grew faster than offspring from high-condition mothers. However, such differences were only observed when the offspring were fed high food; no differences in growth were found on low food. In other cases, the evidence for parental effects was largely absent. Vorburger et al. (2008) showed that parental exposure to parasitoids in the aphid (Myzus persicae) yielded maternal effects that were weak or absent.

In addition to qualitatively summarizing the evidence for condition transfer effects, we also examined the magnitude of condition transfer effects. For any given trait that provided evidence for a significant parental effect, we calculated the magnitude of such an effect for the offspring generation exposed to low-quality conditions via the following formula: absolute value [(offspring of parent from low-quality treatment–offspring of parent from high-quality environment)/average offspring trait value] (see Table S1). These data were extracted using the figure calibration plugin for ImageJ (NIH). For all of the cases that provided evidence for a significant condition transfer effect, the offspring of parents that experienced high-quality conditions exhibited trait values that should enhance fitness compared with the offspring of parents that were exposed to low-quality conditions (i.e. faster development, larger body size, faster growth etc.). The same is true for the studies that yielded significant evidence for anticipatory effects. As a result, this quantitative approach is a proxy for the magnitude of the fitness advantages associated with distinct forms of parental effects.

This approach yielded strong effects associated with anticipatory and condition transfer effects (Fig. 2). The average treatment effects in the offspring generation were 32% and 41% for condition transfer and anticipatory effects, respectively (Fig. 2A). This means that for condition transfer effects the average difference in offspring trait values in the low-quality environment was 32% between the offspring from parents that experienced high versus low-quality conditions. Because resource-related manipulations were the most common treatment utilized by the studies included in this Review (Table 1) and also because resources are likely important to condition transfer effects, we also evaluated these same trends for studies that manipulated resources. This broadly included studies that altered resource levels, resource quality, density and water availability for plants. The trends were similar (see Fig. 2B). It is important to note that we quantified the magnitude of parental effects only for results that provided an example (based upon non-overlapping error bars) of anticipatory (or condition transfer) effects in a subset of the offspring treatments (offspring in the low-quality treatments). Our results show that when studies revealed evidence for a significant parental effect (whether it be condition transfer or an anticipatory effect), the magnitude of the effect was strong. However, the vast majority of studies quantify maternal effects for several traits and did not find evidence for anticipatory (condition transfer) effects for all or even the majority of traits. Thus, the average magnitude of any effect would be weaker if the traits that yielded negative results were also included in our results. Given that Uller et al. (2013) performed a meta-analysis that included all results, this can help to explain why they observed overall weak evidence for anticipatory maternal effects.

We examined the papers that are included in this study to determine if there are commonalities that will predict when and why we might observe condition transfer effects. One immediate trend that is apparent is that condition-dependent transfer effects are common in studies that manipulate parental and offspring resource availability and/or resource quality (also water availability in plants) (Table 1). For instance, of the 29 papers that provided some evidence for condition transfer effects, >75% (22/29) of them directly or indirectly modified resources. This is not necessarily surprising given that the amount or quality of resources experienced by parents will directly influence their ability to provision offspring. Thus, variation in resource levels or quality appears to be a key determinant of the expression of condition transfer effects.

In addition to the type of stressor that influences the expression of maternal effects, we also considered whether there are specific early life advantages that best predict the occurrence of condition transfer effects. In particular, are condition transfer effects simply a byproduct of differences in egg or offspring size? Does parental exposure to low-quality or stressful conditions lead to the production of smaller offspring (and vice versa for high-condition parents)? This does not appear to be the case. While there are examples of high-condition mothers producing larger eggs or offspring (Kyneb and Toft, 2006; Bonduriansky and Head, 2007; Or and Ward, 2007; Matesanz et al., 2022), there are also examples showing that low-condition mothers produce larger eggs/offspring (Beyer and Hambright, 2017; Zhou and Declerck, 2020). Other studies that provided evidence for condition transfer effects did not find any connection between maternal condition and egg or offspring size (Fernández-González et al., 2011; Ehlers et al., 2018; Radersma et al., 2018). It is thus clear that condition transfer effects can provide a link between parental condition and offspring fitness irrespective of an influence on egg or offspring size per se.

The empirical studies often varied in the timing and/or duration of exposure to an environmental stressor. For example, some studies reared parents/offspring in the presence and absence of the stressor for their entire lives while others exposed organisms for shorter periods of time or during discrete life history phases. We evaluated the studies included in this review and listed whether the environmental stressor was imposed when the parents were juveniles versus adults (see Table 1). One apparent trend is that the vast majority of studies that detected condition transfer effects exposed the parental generation to the environmental stressor early in life (as juveniles) (Table 1; 86% or 25/29 of the studies). It is intuitive that the timing and duration of exposure may influence the ability to detect condition transfer effects as exposure to a stressful condition early in life may increase the differences in condition between the low- and high-quality environmental conditions. However, it is also important to note that most studies we consider in this Review exposed juveniles to the stressor in the parental generation (as opposed to the adult stage) (43/55 studies). Thus, the relative importance of exposure early versus later in life in terms of the induction of parental effects is less clear based upon the studies we included.

We have largely focused on condition-dependent transfer effects in a laboratory setting. However, it is important to emphasize that condition transfer effects have been observed in captive populations (Parker, 2002), in correlational studies in natural populations (Coulson and Porter, 1985; Ciconia, 2001; Van de Pol et al., 2006) and in field experiments (Gagliano and McCormick, 2007; Allen et al., 2008; Giordano et al., 2014; Burton et al., 2020). For example, increased body condition is associated with increased clutch size in birds (Ciconia, 2001). High parental condition can also have long term fitness consequences. In a 20-year mark–recapture study, developing in a high-quality habitat and increased parental condition was shown to significantly increase lifetime reproductive success in oystercatchers (Haematopus ostralegus) (Van de Pol et al., 2006). In an experimental study on a coral reef fish (Pomacentrus amboinensis), Gagliano and McCormick (2007) showed that high-condition females (due to food supplementation) provisioned eggs with increased energy reserves. Eggs with larger yolk sacs were more likely to hatch and the subsequent offspring exhibited higher survival. There is thus good evidence that condition transfer effects occur in more natural situations.

Our Review argues that condition transfer effects are indeed more widespread than is currently appreciated. This supports the predictions of previous researchers regarding the underrealized importance of condition transfer effects (Bonduriansky and Crean, 2018). We found evidence for condition transfer effects across a variety of taxa and environmental stressors (Table 1). Although there are studies that document condition transfer effects in response to several environmental stressors (Table 1), the majority of examples of condition transfer effects involved parental manipulations of resource availability or resource quality (or water availability for plants) (Table 1). Such results signal that condition transfer effects may be less likely to be driven by external environmental signals but could instead be more strongly determined by factors that dictate parental condition. This trend is not particularly surprising given that condition transfer effects rely upon females or parents differing in condition which should be linked to the resources that they experience during development. With that said, we feel there are several outstanding avenues of future research.

Our Review collated information on whether one or both parents were exposed to the environmental stressors (Table 1). This is important because offspring development can be influenced by environmental effects on mothers as well as fathers (Bonduriansky and Head, 2007; Zizzari et al., 2016; Tariel et al., 2020). For example, the parental influence on offspring escape behavior in freshwater snail (Physa acuta) depended upon whether mothers or fathers were exposed to predator cues (Tariel et al., 2020). Such a result is important because we included just three papers that utilized experimental designs that could quantify the relative importance of maternal versus paternal effects on offspring fitness (Table 1). As a result, an increased understanding of the relative contributions of maternal versus paternal effects in the context of transgenerational plasticity should be a priority for future research.

Condition transfer effects do not appear to manifest because of a simple connection between maternal resource availability during development and shifts in egg or offspring size (see above). Several studies explicitly made the point that offspring size does not predict offspring fitness (see Beyer and Hambright, 2017). Increased attention to how and why differences in maternal or paternal condition yields differences in offspring fitness is warranted. This includes a better understanding of the underlying hormonal, physiological, and non-genetic mechanisms of condition transfer effects.

Our Review provided some evidence that the duration of exposure may be important to the induction of condition transfer effects (see above). The majority of studies that provided evidence for condition transfer effects manipulated the environmental stressor when parents were young. Yet, this approach was not universal (Table 1). More work is needed to better understand when condition transfer effects are likely to manifest in wild populations. In particular, a better understanding of how the timing, magnitude and duration of environmental stressors induce condition transfer effects is needed.

The studies included in this Review were skewed towards the use of invertebrates as study organisms (34/55 studies, 62%). This likely stems from the experimental advantages of using organisms that are small and display a short generation time when conducting multi-generation experimental studies. This Review included just 10 studies that used vertebrate species. Furthermore, most of these studies on vertebrates exposed the parental generation to an environmental stressor when they were adults (8/10 studies). Experimental studies that examine parental effects in vertebrates when such organisms experience low-quality conditions early in life or throughout development are lacking and should be a priority for future research.

We provided evidence that condition transfer exists and may influence fitness in natural populations (Coulson and Porter, 1985; Ciconia, 2001; Van de Pol et al., 2006). However, more experiments in nature across a broader range of taxa would be beneficial to determine the broad ecological importance of condition transfer effects.

A small number of studies provided evidence for interpopulation or interspecific differences in the degree or forms of maternal effects (e.g. Bashey, 2006; Sultan et al., 2009). In these examples, the populations/species originated from contrasting environments. Such results foreshadow that there may be specific ecological conditions that favor the evolution of condition transfer effects. As a result, more work is needed from both a theoretical and empirical perspective to determine the ecological conditions that drive the evolution of condition transfer effects.

M.R.W. thanks Katie Gilmour and Pat Wright for the invitation to the JEB 2023 symposium and for the Company of Biologists for providing funding for the symposium. M.R.W. would also like to thank all of the participants for stimulating presentations and discussions.