Limnol. Oceanogr. 66, 2021, 201–213 © 2020 The Authors. Limnology and Oceanography published by Wiley Periodicals LLC on behalf of Association for the Sciences of Limnology and Oceanography. doi: 10.1002/lno.11597 The fecal iron pump: Global impact of animals on the iron stoichiometry of marine sinking particles Priscilla K. Le Mézo ,1,2* Eric D. Galbraith1,3,4 1 Institut de Ciència i Tecnologia Ambientals (ICTA), Universitat Autonoma de Barcelona (UAB), Barcelona, Spain 2 Laboratoire de Météorologie Dynamique (LMD) / Institut Pierre Simon Laplace, CNRS, Ecole Normale Supérieure, Université PSL, Ecole Polytechnique, Sorbonne Université, Paris, France 3 Catalan Institution for Research and Advanced Studies (ICREA), Barcelona, Spain 4 Earth and Planetary Sciences, McGill University, Montreal, Quebec, Canada Abstract The impact of marine animals on the iron (Fe) cycle has mostly been considered in terms of their role in sup- plying dissolved Fe to phytoplankton at the ocean surface. However, little attention has been paid to how the transformation of ingested food into fecal matter by animals alters the relative Fe-richness of particles, which could have consequences for Fe cycling in the water column and for the food quality of suspended and sinking particles. Here, we compile observations to show that the Fe to carbon (C) ratio (Fe:C) of fecal pellets of various marine animals is consistently enriched compared to their food, often by more than an order of magnitude. We explain this consistent enrichment by the low assimilation rates that have been measured for Fe in animals, together with the respiratory conversion of dietary organic C to excreted dissolved inorganic C. Furthermore, we calculate that this enrichment should cause animal fecal matter to constitute a major fraction of the global sinking flux of biogenic Fe, a component of the marine iron cycle that has been previously unappreciated. We also estimate that this fecal iron pump provides an important source of Fe to marine animals via coprophagy, particularly in the mesopelagic, given that fecal matter Fe:C can be many-fold higher than the Fe:C of local phy- toplankton. Our results imply that the fecal iron pump is important both for global Fe cycling and for the iron nutrition of pelagic and mesopelagic communities. Iron (Fe) is an essential element used to support many role in determining the total primary productivity in Fe- functional aspects of marine organisms, including photosyn- limited regions (Boyd and Ellwood 2010). thesis, cellular respiration, and oxygen transport (Bury and Heterotrophic consumers, including multicellular animals, Grosell 2003a; Marchetti and Maldonado 2016). Its scarcity in have received attention as recyclers of Fe within the euphotic the ocean has been shown to limit phytoplankton growth, zone of Fe-limited regions (e.g., Sarthou et al. 2008; Ratnarajah especially in the nitrate-rich High Nutrient Low Chlorophyll et al. 2018). For example, in an high nutrient, low chlorophyll regions in which the addition of Fe can boost primary produc- region, Sarthou et al. (2008) showed that copepod grazing tion (Moore et al. 2013). Moreover, it has been suggested that increases Fe recycling, which supports about half of the local low Fe concentrations can be limiting to marine animals, phytoplankton demand for Fe, and Nuester et al. (2014) also including zooplankton (Chen et al. 2011) and fish (Galbraith highlighted that dissolved Fe released by grazing meso- et al. 2019). Because of its importance as a trace nutrient, the zooplankton is taken up by phytoplankton faster than inorganic efficiency with which Fe is recycled at the surface, instead of Fe. The activity of consumers, especially zooplankton, therefore being lost to depth in sinking particles, plays an important plays a major role in the recycling of Fe in the surface layer. Yet, marine animals not only recycle Fe within the surface ocean, they also produce Fe-bearing particulate fecal matter (i.e., fecal pellets), with overall consequences for the iron cycle *Correspondence: priscilla.le-mezo@lmd.ens.fr that remain poorly defined. On one hand, the export of Fe by This is an open access article under the terms of the Creative Commons fecal pellets can contribute to Fe limitation in the surface Attribution License, which permits use, distribution and reproduction in ocean due to their large sinking speed (Schmidt et al. 1999) any medium, provided the original work is properly cited. and sometimes refractory nature (Cabanes et al. 2017). But on Additional Supporting Information may be found in the online version of the other hand, prior work has argued that fecal matter fertil- this article. izes the water with Fe when it is remineralized (Schmidt 201 Le Mézo and Galbraith The fecal iron pump et al. 2016; Laglera et al. 2017), which might suggest an equiv- and consuming sinking fecal matter will be strongly influenced ocal overall outcome on Fe recycling. Thus, although sinking by the stoichiometry of this fecal matter. Interestingly, there is a fecal pellets can contribute to a significant part of the export relatively high abundance of mesopelagic fish observed in Fe- of organic carbon (e.g., Turner 2015) and can have an Fe:C limited high nutrient, low chlorophyll regions (e.g., Beamish that is significantly different than other particles et al. 1999; Moteki et al. 2011), where epipelagic fish are scarce (e.g., Cabanes et al. 2017), few studies have considered the (Galbraith et al. 2019), prompting the question of whether or role of animals in the particulate Fe distribution in the water not fecal pellets might alleviate Fe limitation among mesopelagic column (an exception being the observational study of Laglera organisms compared to epipelagic ones. et al. 2017). Furthermore, little consideration has been given In this article, we develop the hypothesis that Fe:C is con- to how animals might contribute to the ways in which the sistently enriched in the fecal matter of animals, relative to the food they ingest, and that this plays an important role in vertical distribution of particulate Fe differs from that of C, setting the vertical distribution of biogenic Fe in the water col- nitrogen, and phosphorus in the global ocean. umn. We refer to this overall process as the “fecal iron pump.” One distinctive feature of the vertical distribution of Fe is We additionally propose that this fecal iron pump provides an that its total concentration in particles tends to increase with important source of Fe nutrition to organisms, especially in depth. The Fe:C of sinking matter, determined from the sink- the mesopelagic community. We provide an initial test of ing flux of particulate Fe vs. the sinking flux of particulate C, these hypotheses by compiling Fe:C measurements in feces as also generally increases with depth (e.g., Frew et al. 2006; well as data on Fe:C in marine organisms and on Fe and C Bressac et al. 2019). These observed enrichments of Fe in sink- absorption and assimilation efficiencies. We also build a sim- ing particles are typically attributed to abiological process, pri- ple model to provide rough estimates of the global impact of marily the scavenging of Fe onto sinking particles, sediment fecal Fe-enrichment on sinking particles in the ocean. Our resuspension and/or water mass mixing. Fecal pellets are not results support a significant role for the fecal iron pump, both widely considered to contribute to the vertical changes in par- as an overlooked part of the global Fe cycle and as a source of ticulate Fe:C due, at least in part, to the lack of a general this critical micronutrient to the mesopelagic community. expectation for how the Fe content of feces should differ from the Fe of animal food. Fecal matter also serves as a food source for many organ- Methods isms in the whole water column. (Bailey and Robertson 1982; Conceptual basis for data analysis Köster and Paffenhöfer 2017), but is proportionally more In order to analyze diverse data sources within a common important in the mesopelagic and bathypelagic layers where framework, we first developed a framework within which to coprophagy makes up a larger part of the overall diet quantify the fates of ingested Fe and C in a generalized con- (e.g., González and Smetacek 1994; Sampei et al. 2009). Thus, sumer, illustrated in Fig. 1. Part of the ingested food is the nutritional status of mesopelagic organisms living at depth absorbed through the gut epithelium, as determined by the absorption efficiency (also referred to elsewhere as the gut assimilation efficiency): we call AX the absorption efficiency of an element X (Fig. 1). The food that is not absorbed is egested as fecal matter, (1 − AX)IX where IX is the quantity of X ingested. Part of the absorbed food is allocated to the con- struction of new biomass (including somatic growth and reproductive tissue), αXAXIX where αX is the somatic assimila- tion efficiency of an element X. The remainder is excreted, given by (1 − αX)AXIX. We call τX the trophic efficiency of an element X, so that the amount of element X incorporated in new matter is τXIX and τX = AXαX. Note that the trophic effi- ciency of C is sometimes called the gross growth efficiency. In this framework, the ingested fraction that we consider is the one that actually enters the digestive tract. We note that sloppy feeding, by which elements can be lost before entering the digestive tract as represented on Fig. 1 may also affect the Fig 1. Simple representation of Fe:C stoichiometry transformations arising Fe:C of the food and thus the measurements of efficiencies. from the passage of food through a marine animal with a gut. The arrows For example, Møller et al. (2003) estimated for copepods that represent the fluxes of elements. The colored boxes represent the Fe:C stoichi- ometry in the different components of the system. τX is the trophic efficiency about 50% of particulate C is lost from food by sloppy feeding. for element X and AX is the absorption efficiency for element X. Sloppy feed- We are not aware of any study quantifying the amount of Fe ing releases dissolved organic Fe and C, DFe and DOC, respectively. lost from the food via sloppy feeding and therefore assume 202 Le Mézo and Galbraith The fecal iron pump that Fe and C are released in similar amounts during feeding, Calculation of the Fe:C enrichment of sinking particles due similar to Laglera et al. (2020). to fecal pellets Also, it has been shown that Fe can be taken up directly The Fe:C enrichment of the particulate organic pool, Epart, from water by fish via their gills, but this direct absorption is that occurs for a given fraction of fecal pellets in the sinking negligible at the low dissolved Fe concentrations found in sea- flux, x, depends on the Fe:C enrichment of the fecal pellets water (Bury and Grosell 2003b). We therefore consider food as compared to the Fe:C in food, EFP as: Epart = xEFP + (1 − x), the only source of Fe to marine animals. with x the fraction of particulate organic carbon (POC) that is fecal pellets. The contribution of fecal pellets Fe to the total Fe in the Data on absorption and assimilation efficiencies of particulate pool is thus: Fe and C ðFe:CÞFP :CFP ðFe:CÞ ðFe:CÞ Table 1 shows data gathered from the literature on Fe and FeFP = ðFe:CÞ = ðFe:CÞ FP  ðFe:CÞfood  x = EEpart FP  x , with CFP the Fepart part : POC food part C absorption efficiencies for different organisms, and data on fecal pellet carbon content; (Fe:C)part, (Fe:C)FP, and (Fe:C)food measured Fe assimilation or trophic efficiencies. For C assimi- the Fe:C in particles, fecal pellets, and food, respectively. lation efficiency, we refer to Sterner and Elser (2002) who give This simple computation is made assuming the Fe:C in non- a C assimilation efficiency of 10–20% for marine animals. fecal organic particles is equal to the Fe:C of the animal’s Some of the measurements of absorption and assimilation food, which ignores the fact that the nonfecal organic parti- efficiencies listed in Table 1 may be biased by sloppy feeding, cles are composed of more diverse material and subject to as mentioned above. The experimental values for which degradation by free-living bacteria. These estimates, shown sloppy feeding may have altered the determination of absorp- in Fig. 3, should be seen only as a rough illustration, rather tion or assimilation efficiencies are indicated in italic in than a precise quantification of the effect on the particulate Table 1. However, one can note that the assimilation effi- pool Fe:C. ciency may also represents the overall net result of all the pro- cesses that start from untouched food and produce feces and new matter, including sloppy feeding and absorption across Calculation of the Fe nutrition of coprophages the gut. To illustrate how the Fe:C enrichment in fecal matter could supplement the Fe nutrition of coprophagous hetero- Measured Fe:C in fecal matter vs. food trophs, we calculate the consequences of variable Fe and C To test the general expectation that Fe is enriched in fecal stoichiometry in a simple coprophagous food web (Fig. 4a). matter, we gathered data on the Fe:C in fecal matter and Fe:C We assume that organisms have constant absorption and in food. We divide the data into three categories (paired, assimilation efficiencies for Fe of 0.3 and 0.2, respectively, stomach-paired, and unpaired) depending on the availability and for C of 0.8 and 0.2, respectively (estimated from of measurements of the Fe:C of food associated with a given Table 1). We define three degrees of coprophagy depending fecal matter Fe:C. We define the preferred “paired” category as on the dietary fraction of fecal matter and on the feeding feeding experiments and experiments where the food was col- behavior of the organism that produced the fecal matter. A lected at the same time and location as the fecal matter. For coprophagy level of 0 indicates a purely planktivorous diet, the “stomach-paired” measurements, we use the stomach con- a coprophagy level of 1 indicates feeding on a mixture of tent Fe:C as an approximation of the Fe:C in the food ingested plankton and fecal matter produced by an organism only (Geesey et al. 1984). The “unpaired” measurements, which are eating primary producers, and a coprophagy level of 2 indi- the least reliable, use the identified prey mean Fe:C, when cates feeding on a mixture of plankton and fecal matter pro- referred to in the same study, or the stomach content Fe:C duced by an organism of coprophagy level 1 (Fig. 4a). Thus, from other studies with the same animal (details and refer- the Fe:C of an organism of a coprophagy level i is given by: ences in Supporting Information Table S1). (Fe:C)body,i = coproi * (Fe:C)FP,i−1 + (1 − coproi) * (Fe:C)plankton, In addition to organic matter, marine animals and their where coproi is the percentage of the diet composed of fecal prey can ingest lithogenic particles from which iron can be pellets, (Fe:C)FP,i−1 is the Fe:C of the fecal pellets produced extracted and absorbed (e.g., Maranger et al. 1998; Schmidt by the organisms of coprophagy level i − 1, and (Fe:C)plankton is et al. 2011, 2016). Thus, when using ranges of Fe:C in identi- the Fe:C of plankton. fied prey without including the ingestion of lithogenic or These feeding behaviors, simplified here, have been detrital particles, we may be overestimating some of the observed for instance in coral reef fishes at Palau in the West- enrichment factors. Additionally, uncertainty on the Fe:C of ern Pacific (Bailey and Robertson 1982). We show several sce- the food that is actually ingested may also arise from the narios in which we vary the fraction of coprophagy to span release of Fe and C via sloppy feeding as discussed earlier. All the range of observed coprophagic behaviors: 5%, 30%, and the ratios of fecal matter to food Fe:C are represented on Fig. 2 100% of the diet (Fig. 4a) (Frankenberg and Smith 1967; and listed in Supporting Information Table S1. González and Smetacek 1994; Iversen and Poulsen 2007). 203 Table 1. Absorption efficiencies of Fe and C, AFe and AC, and assimilation (or trophic) efficiency of Fe, τFe, for marine plankton and fish fed different diets. Maxi- mum efficiencies equal 1. “Fe-poor” and “Fe-rich” indicate whether the food was cultivated under low or high Fe concentrations. Values in italic may have been biased by sloppy feeding. Organism Food AFe AC τ Fe Reference Calanoid copepod assemblage, mostly Acartia Fe-poor Thalassiosira weissflogii 0.17 0.46 — Schmidt et al. 1999 spp. Fe-rich Thalassiosira weissflogii 0.10 0.44 Le Mézo and Galbraith — Calanoid copepods assemblage Heterotrophic ciliate Uronema sp. 0.32±0.02 — — Twining and Fisher 2004 Calanoid copepods assemblage Heterotrophic dinoflagellate 0.66±0.06 — — Oxyrrhis marina Copepod Acartia tonsa Thalassiosira oceanica + Fe-poor 0.53 0.89 — Chen et al. 2014 Rhodomonas salina + Fe-poor 0.43 0.82 — Isochrysis galbana + Fe-poor 0.69 0.87 — Thalassiosira oceanica + Fe-rich 0.47 0.90 — Rhodomonas salina + Fe-rich 0.45 0.87 — Isochrysis galbana + Fe-rich 0.50 0.84 — Mixed copepods assemblage Diatoms + microzooplankton <0.30 — 0.13–0.25 Sarthou et al. 2008 Brine shrimp Artemia sp. Diatom — — 0.19 Hutchins and Bruland 1994 Copepod (unidentified) Diatom — — 0.20 Copepod Acartia tonsa Diatom — — 0.07 Copepod Acartia tonsa Flagellate — — 0.18 Copepod Acartia tonsa + cladoceran Evadne Diatom — — 0.22 204 sp. Flagellate — — 0.10 Calanoid copepod Acartia tonsa Diatom Thalassiosira pseudonana — 0.25±0.03 Hutchins et al. 1995 Diatom Thalassiosira pseudonana — 0.22±0.05 Calanoid copepods assemblage Diatoms Thalassiosira pseudonana — 0.84 — Reinfelder and Fisher 1991 Copepod Calanus pacificus (adult female) Diatom Thalassiosira weissflogii — 0.78±0.07 (0.69–0.85) — Landry et al. 1984 Chaetognath Sagitta hispida Copepods 0.80 (0.54–0.97) Cosper and Reeve 1975 Atlantic salmon, Salmo salar Fish meal or fish meal + wheat 0.11–0.14 — — Storebakken et al. 2000 gluten diets Atlantic salmon, Salmo salar Feed with fish meal + blood meal 0.11–0.33 (0.24±0.04) — — Thodesen et al. 2001 + wheat + fish oil Marine medaka Larval fish Diets with different Fe — — 0.016–0.185 Wang and Wang 2016 Oryzias melastigma Adult female supplements — — 0.009–0.062 Common sole Commercial pellets with different 0.14–0.17 — — Kals et al. 2016 Fe sources (haem/non haem Fe) Damselfish Chromis chromis Plankton (mostly small copepods) 0.84 Pinnegar et al. 2007 Larval Leiostomus xanthurus Lacépède Rotifer Brachionus plicatilis + wild 0.67–0.99 Govoni et al. 1982 plankton The fecal iron pump Le Mézo and Galbraith The fecal iron pump and be repackaged into fecal matter while most of the C is absorbed across the gut wall, respired, and excreted to the dis- solved phase. The overall expectation is thus for a significant enrichment of the Fe:C to occur in fecal matter compared to the food ingested. Note that in the framework we described in the “Methods” section, there is no effect of sloppy feeding since we only con- sider the food that enter the digestive tract. However, since some of the measurements in Table 1 might be biased by sloppy feed- ing, it would affect the description of the processes such that C is not solely lost via excretion but also during ingestion by sloppy feeding along with some losses of Fe, which does not change our expectations for enriched fecal matter. Measured Fe:C in fecal matter vs. food Our compilation of measured feces/food pairs confirms the expected enrichment fecal matter Fe:C compared to the ingested food, with all means of the enrichment factors greater than one (Fig. 2). The enrichment of fecal matter Fe:C in paired or stomach-paired measurements is as low as a factor Fig 2. Feces enrichment factor for zooplankton (zoo), fish, and whales. of 1.3 for copepods fed cultured diatoms (Schmidt et al. 1999) Values are categorized between paired measurements (green points), that and up to 55 for the fresh feces of blacksmith fish (Geesey is, when in the study Fe content of feces and the food ingested were both et al. 1984). Old fecal matter of fish exhibits the highest measured, stomach-paired measurements (orange points), that is, when enrichment values up to 120 for the oldest fecal material (130- food Fe was not directly measured but undigested stomach content was used instead and, unpaired measurements (gray points), that is, when the days old). When averaged across all pairs (excluding old fecal food Fe content measurement is not directly related to the food ingested matter and range-only values), Fe:C in fecal matter is larger and may come from one or several other studies. Values given without than Fe:C in the corresponding food by an order of magnitude the measurement uncertainty of the feces Fe:C and/or the food Fe:C are (geometric mean of 9.0 ± 3.3). An order of magnitude enrich- plotted as diamonds. The horizontal gray lines with no points are based ment would be consistent with an average Fe absorption of on a range of values for either Fe:C in feces or food or for both. The verti- cal red line represents equal Fe:C in feces and food, i.e., no enrichment. 10% and C absorption of 90%. Notably, even though the organisms included in Fig. 1 have different digestive systems, exploit various food sources, and handle their food in differ- ent ways (e.g., filterers, mechanical crushing of the prey), their Results and discussion fecal matter appears to be universally enriched in Fe:C com- Observed enrichment of Fe in feces pared to their food (Fig. 2).The variability in observed ratios The contrasting fates of Fe and C in animal guts may be explained by a combination of factors, including (1) Table 1 shows published data for different organisms on uncertainty in the Fe:C of the corresponding food source, the absorption efficiencies of Fe and C, as well as trophic effi- which can be linked to sloppy feeding as discussed earlier, ciencies of Fe. In general, the absorption of carbon is high, (2) differences in the elemental absorption efficiencies while the absorption of Fe is low (AFe < AC). In contrast, the between animals, and (3) post-egestion changes in the Fe and trophic efficiency of C is similar to that of Fe (τC ≈ τFe). Thus, C contents as highlighted by changes observed in the aging of while the absorption efficiency of C is high, resulting in most fish fecal matter. The next section discusses the latter two fac- of the absorbed C being excreted as dissolved compounds tors in greater depth. (τC < AC, Table 1), the trophic efficiency of Fe is similar to the Fe absorption efficiency (τFe ≈ AFe, Table 1), implying that Potential sources of variability in fecal Fe:C most of the absorbed Fe is assimilated. Essentially, this basic Variability in Fe absorption contrast between the elements reflects the fact that most of Marine animals can vary their uptake of Fe to some degree the ingested C is respired by the animals to support metabolic in order to avoid excess Fe. At high concentrations, Fe is toxic needs while there is little difference between the amount of Fe and can favor pathogen growth and impair development absorbed and the amount retained within the body, that is, (e.g., Grosell et al. 2011). While basal losses of Fe occur, no only a small fraction of absorbed Fe is excreted back to water regulated excretory process has been clearly identified as dissolved Fe (Fig. 1; Table 1). (Andersen et al. 1997; Bury and Grosell 2003a). It would there- The assimilation data therefore show that, as illustrated in fore appear that animals, including humans, can regulate their Fe Fig. 1, most of the Fe ingested should transit through the gut uptake to avoid toxicity (Papanikolaou and Pantopoulos 2005), 205 Le Mézo and Galbraith The fecal iron pump 60 but that Fe absorption is limited to a relatively low fraction of the (Fe:C) enrichment factor of the organic particulate pool total due to challenges related to Fe chemistry. We do not have a due to fecal pellets complete understanding of the reasons why the absorption of Fe E FP tends to be low, but we can identify a number of relevant factors affecting absorption that are likely to vary between organisms. While all organisms use enzymes during digestion, either 20 (99% confidence interval) produced by the animals or by its gut microbiome, to break 10 down complexes and extract essential elements from the food 15 E part (Donachie and Zdanowski 1998; Grosell et al. 2011; Freese et al. 2012), some animals also possess gastric glands and can 10 operate an acid lysis of the food cells to release their content (Grosell et al. 2011; Štrus et al. 2019). In addition, upon inges- 5 tion some animals are able to mechanically release the Fe from their preys by crushing them in their foregut and/or triturat- ing the food in their gut (Grosell et al. 2011; Štrus et al. 2019). Once Fe is released from the food, it must be prevented 1 from precipitating and absorbed across the gut epithelium, 0 20 40 60 80 100 which can be difficult depending on the gut environment. For Fecal matter contribution to POC export (x,%) example, the osmotic regulation of fish produces bicarbonate Fig 3. Theoretical estimate of the total Fe:C enrichment factor of the that binds to Fe and precipitates in the form of iron carbon- organic fraction of particles as a function of fecal matter contribution to ates, thus increasing the difficulty of extracting iron from the the POC export (x-axis) and as a function of the fecal matter Fe:C enrich- food (Bury and Grosell 2003a; Grosell et al. 2011). Other ele- ment factor (color). The variable Fe:C enrichment factor of the fecal mat- ter compared to food is based on data from Fig. 2 (excluding old fecal ments present in the food such as copper may interfere with matter) and varies around its mean value (black line on the color bar) at the absorption of Fe as it uses the same route of absorption the 99% confidence interval. The colored area on the plot shows the total (Kals et al. 2016), while others such as ascorbic acid may particulate organic pool enrichment using the variable enrichment factor enhance Fe transfer to the gut epithelium (Cooper et al. 2006). of fecal matter. The black and gray lines on the plot show the total partic- A low pH in the gut of animals with gastric glands reduces the ulate organic pool enrichment factor trajectory using the mean and 95% confidence interval of the fecal matter enrichment factor, respectively. oxidation rate of Fe2+ (Liu and Millero 2002), and more impor- tantly the presence of low oxygen zones in the gut (e.g., for (a) (b) Fig 4. Simple model of how the Fe:C of heterotrophic consumers body (Het) and fecal pellets (FP) could be enriched by coprophagy (coprophagous ani- mals both in orange). Panel (a) shows the feeding behaviors of the different heterotrophs (Het). The organism can be of three levels of coprophagy: 0 = not eating any fecal matter (Het0), 1 = eating fecal matter from a 0-level coprophagy organism plus plankton (Het1), 2 = eating fecal matter from a level 1 coprophagy organism plus plankton (Het2). The fraction of fecal matter included in the diet, f, is constant for all coprophages, and is shown for fractions of 5%, 30%, and 100%. The assimilation efficiencies of C and Fe are taken equal to 0.2 and the absorption efficiencies are 0.3 and 0.8 for Fe and C, respectively. The Fe:C in the plankton compartment is fixed. Panel (b) shows the enrichment factors of the Fe:C of the body (filled bars) and the fecal pellets (dashed bars) of the different organisms, under the three scenarios and for the different coprophagy levels. 206 Le Mézo and Galbraith The fecal iron pump copepods, Tang et al. 2011) forces the reduction of Fe3+ to fecal pellets, is maintained. The presence of a peritrophic Fe2+, the latter of which is more bioavailable and binds to spe- membrane that maintains the integrity of the fecal pellets and cific proteins in order to cross the gut epithelium of fish for favors their export to depth is likely limiting exchanges with example (Cooper et al. 2006). In addition, secretions such as the environment (Frangoulis et al. 2004), when it is not mucus in fish gut that binds to Fe2+ and Fe3+, and the presence degraded (Lautenschlager et al. 1978) or eaten by organisms of ligands, either ingested or produced by the gut microbiome, (Lampitt et al. 1990). Without any perturbations, Hutchins help maintain Fe in solution (Tortell et al. 1999; Cooper et al. (1995) showed that the copepod fecal pellets retained et al. 2006; Hunter and Boyd 2007). Inversely, ingested patho- most of their Fe (> 80%) over 30 days and similarly, Cabanes gens and nonresident bacteria present in the gut (e.g., Tang et al. (2017) showed that Fe in salp fecal pellets was highly et al. 2010) may compete for Fe, thus reducing its absorption. refractory. If they are instead disrupted by animals, for exam- In addition to variations caused by organism-specific diges- ple, via coprohexy or coprochaly, fecal particles release Fe and tive features, experiments have shown that Fe absorption organic C to the dissolved pool (Turner 2015), and the frag- depends on the form and distribution of Fe within the food. mentation of fecal particles into smaller pieces allows for a For example, the Fe absorption efficiency of copepods more efficient recycling of Fe by bacteria returning Fe more increases linearly with the Fe content of the cytoplasm of their rapidly to the water column (e.g., Noji et al. 1991; Iversen and prey, while Fe bound to the exoskeleton or cell membrane Poulsen 2007). appears more difficult to absorb (Chen et al. 2014). Heme Fe, More important for the stoichiometry of sinking particles that is, bound within Fe-porphyrin complexes found in wide- are interelement differences in the ways in which Fe and C spread metalloproteins including hemoglobin and myoglobin exchange with the environment. The release of organic C (Hogle et al. 2014), is more efficiently absorbed than nonheme from fecal pellets has been shown to be fast due to the pres- Fe by fish (Andersen et al. 1997). In some cases, the absorption ence of soluble carbon compounds (Jumars et al. 1989) and efficiency of Fe has also been shown to vary with the Fe con- more importantly, it is faster than the release of Fe in fecal pel- centration in food, with a higher absorption when the food lets and sinking particles (Hutchins et al. 1995; Frew et al. 2006; had a lower Fe content (Andersen et al. 1997; Schmidt Twining et al. 2014), which should contribute to increasing et al. 1999). Fe:C in fecal pellets over time and with depth. In addition, As a result of the multiplicity of parameters affecting Fe fecal pellets may scavenge dissolved Fe from the water column absorption, different taxa may tend to have different absorp- as they sink to depth, which increases their Fe:C with time, as tion efficiencies, as suggested for example by the more effi- evidenced by Geesey et al. (1984), whose work on fish fecal cient Fe absorption by zooplankton than by fish shown in pellets showed that old feces contained more Fe than fresh Table 1. In addition, Wang and Wang (2016) showed that the ones (Fig. 2). At the same time, this scavenging may be Fe absorption of the marine medaka Oryzias melastigma was counteracted to some degree by the release of ligands from the significantly different between larval and female adult fish, fecal pellets either by the bacteria attached to it or via the dis- thus highlighting a possible effect of age on the regulation of ruption of the pellet (Cabanes et al. 2017; Laglera et al. 2020), Fe absorption, which would also be coherent with the fact which would keep Fe in solution close to the pellet and limit that larvae are building their Fe stores and need to absorb its adsorption. Quantifying the relative importance of these more Fe to meet their requirements compared to adults factors will require future observational studies. (Galbraith et al. 2019). In summary, the low Fe absorption of marine animals rela- The contribution of fecal pellets to the Fe:C of export tive to their C absorption, which is paramount in driving the Fecal pellets have been shown to be an important compo- Fe:C enrichment of feces, arises from a combination of factors nent of the carbon flux in the ocean mostly due to their large ranging from Fe chemical speciation to organism-specific sinking speed, which tends to be positively related to their size ingestion and digestion. and density (Wotton and Malmqvist 2001). The contribution of zooplankton fecal pellets to the sinking POC flux depends Post-egestion modification of fecal matter Fe:C on the time of year, location, rate of pellet degradation, pro- After egestion, the Fe:C of fecal matter is likely to vary as it ductivity, and community composition, and can represent as exchanges with its environment, so that the measured Fe:C of little as < 1% or as much as 100% of the sinking flux of POC fecal matter could differ from the Fe:C it would have had (e.g., Turner 2015; Steinberg and Landry 2017; Belcher when originally expelled. Apart from the inherent fragility or et al. 2019). Fish fecal pellets can also be an important vector resistance of fecal matter (e.g., partly liquid whale feces; of POC in the water column. For example, up to 17% of POC Roman et al. 2014), physical and biological mechanical degra- captured in sediment traps in the Peruvian coastal upwelling dation of the fecal matter may occur and accelerate its was from anchoveta fecal pellets (Staresinic et al. 1983), while recycling, thus modifying its Fe:C. 15–17% of the POC export is estimated to be mediated by The release of both Fe and C from fecal matter is slower mesopelagic fishes (including active transport) off of Southern when the physical integrity of the fecal particles, that is, the California and up to 40% in the North Pacific Subtropical 207 Le Mézo and Galbraith The fecal iron pump Gyre (Davison et al. 2013). Given that fecal pellets are a glob- the mesopelagic and bathypelagic layers (Belcher et al. 2017), ally significant vector for the downward transport of carbon, which would increase the particulate enrichment factor with their high Fe:C implies that they could dominate the down- depth. Reverse fluxes from the seabed to shallow layers by ward transport mechanism for organic Fe in many parts of the organisms feeding on the sediment may also add some vari- ocean. ability in the contribution of fecal matter to the Fe enrich- A rough estimate for how much the presence of Fe-rich ment of particles (Schmidt et al. 2011). fecal pellets could enrich the Fe:C of the organic fraction of Community composition is an important factor in deter- sinking particles is shown in Fig. 3. For example, if fecal pellets mining the contribution of fecal pellets to the sinking fluxes contribute to 30% of the sinking flux of POC, the Fe:C enrich- of organic carbon as it influences the remineralization length ment of the particulate organic pool (Epart) would range scale of the particles. Indeed, the fecal pellets can be mostly between 1.4- and 8-fold based on the 99% confidence interval consumed in the epipelagic, as evidenced for example by the around the geometric mean of the observed Fe:C enrichment presence of copepods that highly increases the recycling of of the fecal pellets (EFP, color-scale on Fig. 3); using the fecal material in the upper layers of the ocean (e.g., Sarthou observed geometric mean of EFP (EFP = 9.0), the mean enrich- et al. 2008; Laglera et al. 2017). In addition to the ingestion of ment of organic particles by the fecal Fe pump would be fecal pellets, that is, coprophagy (González and Smetacek 3.4-fold (Fig. 3). Under this scenario (i.e., a 30% contribution 1994), the fragmentation of fecal pellets (Iversen and of fecal pellets to the sinking flux of POC), between 50% and Poulsen 2007) also reduces their contribution to the sinking flux 91% of the organic sinking flux of Fe would be due to fecal of particles, directly or through faster bacterial remineralization pellets (with a mean value of 80% using the mean due to reduced sinking speeds. Finally, the reprocessing of sink- observed EFP). ing particles at depth through coprophagy also produces new fecal pellets, themselves more enriched in Fe compared to C. Sources of variability in the fecal Fe pump importance The contribution of fecal pellets to the biogenic fraction of Coprophagy and Fe-rich fecal matter particulate Fe is bound to be proportionally larger than the The results in Fig. 4b show the degree to which the Fe:C overall effect on sinking fluxes, due to the presence of enrichment in the body and fecal matter of an organism lithogenic and authigenic Fe fractions. Unfortunately, measur- would be expected to increase with the percentage of ing the biogenic Fe in sinking particles is challenging and the coprophagy and with the coprophagy level, for the illustrative results are methodologically dependent (Rauschenberg and absorption and assimilation rates. The Fe:C increase is greater Twining 2015). Many studies assume that the Fe:P or Fe:C with a higher consumption of fecal matter in the diet, from ratio is constant in biological material, and use the Fe:P or Fe: 1.1-fold (1.1-fold) body-enrichment with 5% coprophagy to C of phytoplankton to compute the Fe biogenic profiles 3.5-fold (12-fold) body-enrichment with 100% coprophagy for (e.g., Bowie et al. 2015; Barrett et al. 2018). Obviously, assum- a coprophagy level 1 (level 2) animal. ing a constant Fe:P or Fe:C ratio will prevent the accurate Reality is sure to be much more nuanced that this simple determination of changes in Fe stoichiometry such as those theoretical calculation indicates, yet the implied Fe:C enrich- identified here. Observed profiles of biogenic particulate Fe ment of fecal matter is quite conservative compared to the worldwide are highly variable, depending on the station and fecal matter Fe:C enrichment values computed from published on the time of the measurements (e.g., Planquette et al. 2011), data (Fig. 2), which shows a mean enrichment of about one adding another difficulty to testing the contribution of marine order of magnitude. These numerical results illustrate the large animals to the Fe:C variations of sinking particulate matter. degree to which the consumption of fecal matter could pro- Complicating the matter further, krill have been shown to vide Fe nutrition to marine animals. ingest lithogenic particles and to extract Fe from them In addition to coprophagy, the ingestion of smaller organ- (Schmidt et al. 2016), and flagellates can incorporate isms that feed on or incorporate the enriched fecal pellets and lithogenic Fe via phagotrophy (Maranger et al. 1998), which is sinking particles via phagotrophy (Maranger et al. 1998; Twin- then transferred to marine animals upon grazing. This sug- ing and Fisher 2004) may also be an important pathway gests that animals are also likely to contribute to the redistri- through which this Fe-rich source food is transferred to bution of particulate lithogenic Fe. Given these complications, marine animals. it is difficult to provide a confident estimate for how the fecal Fe pump compares to the total non-fecal Fe flux. Fe bioavailability in fecal matter Vertical migrations and egestion at depth likely affect the Our data compilations and calculations both show that Fe: importance of fecal pellets in shaping the particulate Fe:C dis- C can be strongly enriched in fecal matter, but the degree to tribution in the water column as these processes can signifi- which the fecal Fe pump may provide an important source of cantly contribute to the POC export (e.g., Boyd et al. 2019). In nutrition depends on the bioavailability of fecal Fe. Some the Southern Ocean, the flux of fecal pellets at depth has been aspects of fecal matter formation would be expected to con- shown to increase due to in situ production of fecal pellets in tribute to low Fe bioavailability, while others should raise the 208 Le Mézo and Galbraith The fecal iron pump bioavailability. On one hand, easily absorbed Fe could be hypothesis could be further tested with new measurements expected to be assimilated first, leaving more refractory Fe to that are able to distinguish changes in the Fe:C of organic be routed to the fecal matter (Hutchins et al. 1995). On the matter within the water column. other hand, Fe in fecal matter can be accessible to other organ- Moreover, we propose that the fecal matter serves as a Fe- isms for absorption. For instance, Schmidt et al. (2016) rich food source to marine animals via coprophagy, especially showed that the proportions of labile iron in krill fecal pellets relevant for the mesopelagic community. This could contrib- was five times higher than in diatoms, and Sarthou et al. (2008) ute to the relatively high abundance of mesopelagic fish in and Laglera et al. (2017) concluded that the recycling of Fe high nutrient, low chlorophyll regions where the pelagic bio- from copepod fecal pellets is responsible for longer phyto- mass of fish seems to be limited by the Fe content of plankton blooms, thus suggesting that Fe in the fecal matter is their prey. bioavailable. Fe bioavailability in fecal matter is affected by The apparent importance of the fecal Fe pump raises many multiple factors such as the presence of undigested cells and questions about how it may respond to anthropogenic organisms (e.g., Aarnio and Bonsdorff 1997; Friedland change. In the context of climate change, rising ocean temper- et al. 2005; Köster et al. 2011), the presence of an acidic phase ature may accelerate fecal matter degradation and/or shift the in the digestion process that can modify Fe speciation and community to smaller animals producing smaller particles help extract Fe from lithogenic particles (e.g., Schmidt (Heneghan et al. 2019), thereby decreasing the fecal Fe pump et al. 2016), the presence of ligands that keep Fe in solution effectiveness. On the other hand, community changes such as and potentially available (e.g., Nuester et al. 2014; Laglera a shift from krill to salps in the Southern Ocean are predicted et al. 2020). In addition, the microbial community within the (Atkinson et al. 2004), which might increase Fe transport to pellet matrix and attached to its surface grows and degrades depth via heavier, poorly digested fecal matter (Cabanes the fecal matter during its sojourn in the water column et al. 2017). At the same time, declining oxygen concentra- (e.g., Tang et al. 2010; Morata and Seuthe 2014) and may alter tions that affect the respiration rates of organisms, and fishing the Fe state within the feces through the release of ligands activity that removes large organisms and modifies ecosystem and low oxygen concentrations within the fecal matter, as in structures and communities (Getzlaff and Oschlies 2017), aggregates (Balzano et al. 2009). Microbes that grow on the might also impact the fecal Fe pump in ways that are chal- fecal matter, attached to the peritrophic membrane, may be lenging to predict. Resolving these questions will require new digested when the fecal matter is ingested (Newell 1965; observations from the field as well as experimental and model- Lampitt et al. 1990; Anderson et al. 2017), potentially consti- ing studies. tuting an additional source of bioavailable Fe. In short, although work is needed to have a clearer under- Acknowledgements standing of the Fe bioavailability of fecal matter to different marine animals, it would appear likely that it is generally a We thank the editors and both reviewers for their helpful rich source of bioavailable Fe. comments. This study was supported by the European Research Council (ERC) under the European Union’s Horizon 2020 Research and Innovation Program (Grant Agreement No Conclusions 682602) and the Spanish Ministry of Science, Innovation and In this article, we have shown that the fecal matter of Universities, through the “María de Maeztu” Program for marine animals is generally enriched in Fe:C compared to Units of Excellence (MDM-2015-0552). their food, primarily because animals digest and respire most of the carbon they ingest, while allowing most of the Fe to pass through and be packaged as fecal matter. This low absorp- References tion efficiency of Fe appears to depend on multiple factors Aarnio, K., and E. Bonsdorff. 1997. Passing the gut of juvenile related to the physiology of the organism and the type of food flounder, Platichthys flesus: Differential survival of it consumes. The digestive enrichment can then be sup- zoobenthic prey species. Mar. Biol. 129: 11–14. http://dx. plemented through the additional processes that preferentially doi.org/10.1007/s002270050140 allow organic C to leak from the fecal pellets, as well as the Andersen, F., M. Lorentzen, R. Waagbø, and A. Maage. 1997. potential scavenging of Fe by the fecal pellets. The sinking of Bioavailability and interactions with other micronutrients this fecal matter therefore enriches the Fe:C of particles at of three dietary iron sources in Atlantic salmon, Salmo salar, depth, a process we term the fecal Fe pump. Given the signifi- smolts. Aquacult. Nutr. 3: 239–246. http://dx.doi.org/10. cant contribution of sinking fecal matter to the export of 1046/j.1365-2095.1997.00096.x organic carbon, combined with its high observed Fe content, Anderson, T. R., D. W. Pond, and D. J. Mayor. 2017. The role we hypothesize that the fecal Fe pump is a dominant pathway of microbes in the nutrition of detritivorous invertebrates: by which biogenic Fe is preferentially exported to depth, rela- A stoichiometric analysis. Front. Microbiol. 7: 1–13. http:// tive to other nutrients, at many places in the ocean. This dx.doi.org/10.3389/fmicb.2016.02113 209 Le Mézo and Galbraith The fecal iron pump Atkinson, A., V. Siegel, E. Pakhomov, and P. Rothery. 2004. Cabanes D. J. E., and others. 2017. First evaluation of the role Long-term decline in krill stock and increase in salps within of Salp fecal pellets on iron biogeochemistry. Front. Mar. the Southern Ocean. Nature 432: 100–103. http://dx.doi. Sci. 3: 1–10. http://dx.doi.org/10.3389/fmars.2016.00289 org/10.1038/nature02996 Chen, X., S. B. Baines, and N. S. Fisher. 2011. Can cope- Bailey, T. G., and D. R. Robertson. 1982. Organic and caloric levels pods be limited by the iron content of their food? of fish feces relative to its consumption by coprophagous reef Limnol. Oceanogr. 56: 451–460. http://dx.doi.org/10. fishes. Mar. Biol. 69: 45–50. http://dx.doi.org/10.1007/ 4319/lo.2011.56.2.0451 BF00396959 Chen, X., N. S. Fisher, and S. B. Baines. 2014. Influence of Balzano, S., P. J. Statham, R. D. Pancost, and J. R. Lloyd. 2009. algal iron content on the assimilation and fate of iron and Role of microbial populations in the release of reduced iron carbon in a marine copepod. Limnol. Oceanogr. 59: to the water column from marine aggregates. Aquat. Microb. 129–140. http://dx.doi.org/10.4319/lo.2014.59.1.0129 Ecol. 54: 291–303. http://dx.doi.org/10.3354/ame01278 Cooper, C. A., N. R. Bury, and M. Grosell. 2006. The effects of Barrett, P. M., J. A. Resing, M. M. Grand, C. I. Measures, and pH and the iron redox state on iron uptake in the intestine W. M. Landing. 2018. Trace element composition of of a marine teleost fish, gulf toadfish (Opsanus beta). Comp. suspended particulate matter along three meridional CLI- Biochem. Physiol. 143: 292–298. http://dx.doi.org/10. VAR sections in the Indian and Southern Oceans: Impact 1016/j.cbpa.2005.11.024 of scavenging on Al distributions. Chem. Geol. 502: 15–28. Cosper, T. C., and M. R. Reeve. 1975. Digestive efficiency of http://dx.doi.org/10.1016/j.chemgeo.2018.06.015 the chaetognath Sagitta hispida. J. Exp. Mar. Biol. Ecol. 17: Beamish, R. J., K. D. Leask, O. A. Ivanov, A. A. Balanov, A. M. 33–38. Orlov, and B. Sinclair. 1999. The ecology, distribution, and Davison, P. C., D. M. Checkley, J. A. Koslow, and J. Barlow. abundance of midwater fishes of the subarctic Pacific gyres. 2013. Carbon export mediated by mesopelagic fishes in the Prog. Oceanogr. 43: 399–442. http://dx.doi.org/10.1016/ Northeast Pacific Ocean. Prog. Oceanogr. 116: 14–30. S0079-6611(99)00017-8 http://dx.doi.org/10.1016/j.pocean.2013.05.013 Belcher, A., C. Manno, P. Ward, S. A. Henson, R. J. Sanders, Donachie, S. P., and M. K. Zdanowski. 1998. Potential diges- and G. A. Tarling. 2017. Copepod faecal pellet transfer tive function of bacteria in krill Euphausia superba stomach. through the meso- and bathypelagic layers in the Southern Aquat. Microb. Ecol. 14: 129–136. http://dx.doi.org/10. Ocean in spring. Biogeosciences 14: 1511–1525. http://dx. 3354/ame014129 doi.org/10.5194/bg-14-1511-2017 Frangoulis, C., E. D. Christou, and J. H. Hecq. 2004. Comparison Belcher, A., and others. 2019. Krill faecal pellets drive hidden of marine copepod outfluxes: Nature, rate, fate and role in pulses of particulate organic carbon in the marginal ice the carbon and nitrogen cycles. Adv. Mar. Biol. 47: 253–309. zone. Nat. Commun 10: 1–8. http://dx.doi.org/10.1038/ http://dx.doi.org/10.1016/S0065-2881(04)47004-7 s41467-019-08847-1 Frankenberg, D., and J. K. L. Smith. 1967. Coprophagy in Bowie, A. R., and others. 2015. Iron budgets for three distinct marine animals. Limnol. Oceanogr. 12: 443–450. biogeochemical sites around the Kerguelen Archipelago Freese, D., T. Kreibich, and B. Niehoff. 2012. Characteristics of (Southern Ocean) during the natural fertilisation study, digestive enzymes of calanoid copepod species from differ- KEOPS-2. Biogeosciences 12: 4421–4445. http://dx.doi.org/ ent latitudes in relation to temperature, pH and food. 10.5194/bg-12-4421-2015 Comp. Biochem. Physiol. B Biochem. Mol. Biol. 162: Boyd, P. W., and M. J. Ellwood. 2010. The biogeochemical 66–72. http://dx.doi.org/10.1016/j.cbpb.2012.04.007 cycle of iron in the ocean. Nat. Geosci. 3: 675–682. http:// Frew R. D., and others. 2006. Particulate iron dynamics during dx.doi.org/10.1038/ngeo964 FeCycle in subantarctic waters southeast of New Zealand. Boyd, P. W., H. Claustre, M. Levy, D. A. Siegel, and T. Weber. Global Biogeochem. Cycles 20: 1–15. http://dx.doi.org/10. 2019. Multi-faceted particle pumps drive carbon sequestra- 1029/2005gb002558 tion in the ocean. Nature 568: 327–335. http://dx.doi.org/ Friedland, K. D., D. W. Ahrenholz, and L. W. Haas. 2005. Via- 10.1038/s41586-019-1098-2 ble gut passage of cyanobacteria through the filter-feeding Bressac, M., and others. 2019. Resupply of mesopelagic dis- fish Atlantic menhaden, Brevoortia tyrannus. J. Plankton solved iron controlled by particulate iron composition. Nat. Res. 27: 715–718. http://dx.doi.org/10.1093/plankt/fbi036 Geosci. 12: 995–1000. http://dx.doi.org/10.1038/s41561- Galbraith, E. D., P. K. Le Mézo, G. Solanes Hernandez, D. Bianchi, 019-0476-6 and D. Kroodsma. 2019. Growth limitation of marine fish by Bury, N. R., and M. Grosell. 2003a. Iron acquisition by teleost low iron availability in the open ocean. Front. Mar. Sci. 6: fish. Comp. Biochem. Physiol. C Toxicol. Pharmacol. 135: 1–13. http://dx.doi.org/10.3389/fmars.2019.00509 97–105. http://dx.doi.org/10.1016/S1532-0456(03)00021-8 Geesey, G., G. Alexander, R. Bray, and A. Miller. 1984. Fish Bury N. R. 2003. Waterborne iron acquisition by a freshwater fecal pellets are a source of minerals for inshore reef com- teleost fish, zebrafish Danio rerio. J. Exp. Biol. 206: munities. Mar. Ecol. Prog. Ser. 15: 19–25. http://dx.doi.org/ 3529–3535. http://dx.doi.org/10.1242/jeb.00584 10.3354/meps015019 210 Le Mézo and Galbraith The fecal iron pump Getzlaff, J., and A. Oschlies. 2017. Pilot study on potential pellet. J. Plankton Res. 33: 1538–1549. http://dx.doi.org/ impacts of fisheries-induced changes in zooplankton mor- 10.1093/plankt/fbr053 tality on marine biogeochemistry. Global Biogeochem. Köster, M., and G.-A. Paffenhöfer. 2017. How efficiently can Cycles 31: 1656–1673. http://dx.doi.org/10.1002/ doliolids (Tunicata, Thaliacea) utilize phytoplankton and 2017GB005721 their own fecal pellets? J. Plankton Res. 39: 305–315. González, H. E., and V. Smetacek. 1994. The possible role of http://dx.doi.org/10.1093/plankt/fbw089 the cyclopoid copepod Oithona in retarding vertical flux of Laglera, L. M., and others. 2017. Iron partitioning during zooplankton faecal material. Mar. Ecol. Prog. Ser. 113: LOHAFEX: Copepod grazing as a major driver for iron 233–246. http://dx.doi.org/10.3354/meps113233 recycling in the Southern Ocean. Mar. Chem. 196: 148–161. Govoni, J. J., D. S. Peters, and J. V. Merriner. 1982. Carbon http://dx.doi.org/10.1016/j.marchem.2017.08.011 assimilation during larval development of the marine tele- Laglera, L. M., A. Tovar-Sanchez, C. F. Sukekava, H. Naik, ost Leiostomus xanthurus Lacépède. J. Exp. Mar. Biol. Ecol. S. W. A. Naqvi, and D. A. Wolf-Gladrow. 2020. Iron organic 64: 287–299. speciation during the LOHAFEX experiment: Iron ligands Grosell, M., A. P. Farrell, and C. J. Brauner [eds.]. 2011. Fish release under biomass control by copepod grazing. J. Mar. physiology: The multifunctional gut of fish, 1st ed. Elsevier. Syst. 207: 103151. http://dx.doi.org/10.1016/j.jmarsys. London. 2019.02.002 Heneghan, R. F., I. A. Hatton, and E. D. Galbraith. 2019. Cli- Lampitt, R. S., T. Noji, and B. von Bodungen. 1990. Marine mate change impacts on marine ecosystems through the biology implications for material flux. Mar. Biol. 104: lens of the size spectrum. Emerg. Top. Life Sci. 3: 233–243. 15–23. http://dx.doi.org/10.1042/ETLS20190042 Landry, M. R., R. P. Hassett, V. Fagerness, J. Downs, and C. J. Hogle, S. L., K. A. Barbeau, and M. Gledhill. 2014. Heme in Lorenzen. 1984. Effect of food acclimation on assimilation the marine environment: From cells to the iron cycle. efficiency of Calanus pacificus. Limnol. Oceanogr. 29: Metallomics 6: 1107–1120. http://dx.doi.org/10.1039/ 361–364. http://dx.doi.org/10.4319/lo.1984.29.2.0361 c4mt00031e Lautenschlager, K. P., N. K. Kaushik, and J. B. Robinson. 1978. Hunter, K. A., and P. W. Boyd. 2007. Iron-binding ligands and The peritrophic membrane and faecal pellets of Gammarus their role in the ocean biogeochemistry of iron. Environ. lacustris limnaeus Smith. Freshw. Biol. 8: 207–211. http:// Chem. 4: 221–232. http://dx.doi.org/10.1071/EN07012 dx.doi.org/10.1111/j.1365-2427.1978.tb01441.x Hutchins, D. A., and K. W. Bruland. 1994. Grazer-mediated Liu, X., and F. J. Millero. 2002. The solubility of iron in seawa- regeneration and assimilation of Fe, Zn and Mn from ter. Mar. Chem. 77: 43–54. http://dx.doi.org/10.1016/ planktonic prey. Mar. Ecol. Prog. Ser. 110: 259–269. http:// S0304-4203(01)00074-3 dx.doi.org/10.3354/meps110259 Maranger, R., D. F. Bird, and N. M. Price. 1998. Iron acquisi- Hutchins, D. A., W.-X. Wang, and N. S. Fisher. 1995. Copepod tion by photosynthetic marine phytoplankton from grazing and the biogeochemical fate of diatom iron. ingested bacteria. Nature 396: 248–251. http://dx.doi.org/ Limnol. Oceanogr. 40: 989–994. http://dx.doi.org/10.4319/ 10.1038/24352 lo.1995.40.5.0989 Marchetti, A., and M. T. Maldonado. 2016. Iron. In M. A. Iversen, M. H., and L. K. Poulsen. 2007. Coprorhexy, coproph- Borowitzka, J. Beardall, and J. A. Raven 233–280. The physi- agy, and coprochaly in the copepods Calanus helgolandicus, ology of microalgae. Springer. Pseudocalanus elongatus, and Oithona similis. Mar. Ecol. Møller, E. F., P. Thor, and T. G. Nielsen. 2003. Production of Prog. Ser. 350: 79–89. http://dx.doi.org/10.3354/ DOC by Calanus finmarchicus, C. glacialis and C. hyperboreus meps07095 through sloppy feeding and leakage from fecal pellets. Mar. Jumars, P. A., D. L. Penry, J. A. Baross, M. J. Perry, and B. W. Ecol. Prog. Ser. 262: 185–191. http://dx.doi.org/10.3354/ Frost. 1989. Closing the microbial loop: Dissolved carbon meps262185 pathway to heterotrophic bacteria from incomplete inges- Moore, C. M., and others. 2013. Processes and patterns of oce- tion, digestion and absorption in animals. Deep-Sea Res. A anic nutrient limitation. Nat. Geosci. 6: 701–710. http://dx. 36: 483–495. http://dx.doi.org/10.1016/0198-0149(89) doi.org/10.1038/ngeo1765 90001-0 Morata, N., and L. Seuthe. 2014. Importance of bacteria and Kals, J., R. J. W. Blonk, H. W. van der Mheen, J. W. Schrama, protozooplankton for faecal pellet degradation. and J. A. J. Verreth. 2016. Effect of different iron sources on Oceanologia 56: 565–581. http://dx.doi.org/10.5697/oc.55- the alleviation of nutritional anaemia in common sole 3.565 (Solea solea). Aquaculture 451: 266–271. http://dx.doi.org/ Moteki, M., P. Koubbi, P. Pruvost, E. Tavernier, and P. A. Hulley. 10.1016/j.aquaculture.2015.08.036 2011. Spatial distribution of pelagic fish off Adélie and George Köster, M., R. Sietmann, A. Meuche, and G. A. Paffenhöfer. V Land, East Antarctica in the austral summer 2008. Polar Sci. 2011. The ultrastructure of a doliolid and a copepod fecal 5: 211–224. http://dx.doi.org/10.1016/j.polar.2011.04.001 211 Le Mézo and Galbraith The fecal iron pump Newell, R. 1965. The role of detritus in the nutrition of two Schmidt, K., C. Schlosser, A. Atkinson, S. Fielding, H. J. Venables, marine deposit feeders, the prosobranch Hydrobia ulvae and C. M. Waluda, and E. P. Achterberg. 2016. Zooplankton gut the bivalve Macoma balthica. Proc. Zool. Soc. Lond. 144: passage mobilizes lithogenic iron for ocean productivity. 25–45. Curr. Biol. 26: 2667–2673. http://dx.doi.org/10.1016/j.cub. Noji, T. T., K. W. Estep, F. Macintyre, and F. Norrbin. 1991. 2016.07.058 Image analysis of faecal material grazed upon by three spe- Schmidt, M. A., Y. Zhang, and D. A. Hutchins. 1999. Assimila- cies of copepods: Evidence for coprorhexy, coprophagy and tion of Fe and carbon by marine copepods from Fe-limited coprochaly. J. Mar. Biol. Assoc. U. K. 71: 465–480. http:// and Fe-replete diatom prey. J. Plankton Res. 21: dx.doi.org/10.1017/S0025315400051717 1753–1764. http://dx.doi.org/10.1093/plankt/21.9.1753 Nuester, J., S. Shema, A. Vermont, D. M. Fields, and B. S. Staresinic, N., J. Farrington, R. B. Gagosian, C. H. Clifford, and Twining. 2014. The regeneration of highly bioavailable iron E. M. Hulburt. 1983. Downward transport of particulate by meso- and microzooplankton. Limnol. Oceanogr. 59: matter in the Peru coastal upwelling: Role of the Ancho- 1399–1409. http://dx.doi.org/10.4319/lo.2014.59.4.1399 veta, Engraulis ringens, p. 225–240. In Coastal upwelling its Papanikolaou, G., and K. Pantopoulos. 2005. Iron metabolism sediment record. Springer. Boston, MA. and toxicity. Toxicol. Appl. Pharmacol. 202: 199–211. Steinberg, D. K., and M. R. Landry. 2017. Zooplankton and http://dx.doi.org/10.1016/j.taap.2004.06.021 the ocean carbon cycle. Ann. Rev. Mar. Sci. 9: 413–444. Pinnegar, J. K., N. V. C. Polunin, J. J. Videler, and J. J. de http://dx.doi.org/10.1146/annurev-marine-010814-015924 Wiljes. 2007. Daily carbon, nitrogen and phosphorus bud- Sterner, R. W., and J. J. Elser. 2002. Ecological stoichiometry: gets for the Mediterranean planktivorous damselfish Chro- The biology of elements from molecules to the biosphere. mis chromis. J. Exp. Mar. Biol. Ecol. 352: 378–391. http:// Princeton Univ. Press. dx.doi.org/10.1016/j.jembe.2007.08.016 Storebakken, T., K. D. Shearer, G. Baeverfjord, B. G. Nielsen, T. Planquette, H., R. R. Sanders, P. J. Statham, P. J. Morris, and Åsgård, T. Scott, and A. De Laporte. 2000. Digestibility of G. R. Fones. 2011. Fluxes of particulate iron from the upper macronutrients, energy and amino acids, absorption of ocean around the Crozet Islands: A naturally iron-fertilized elements and absence of intestinal enteritis in Atlantic environment in the Southern Ocean. Global Biogeochem. salmon, Salmo salar, fed diets with wheat gluten. Aqua- Cycles 25: 1–12. http://dx.doi.org/10.1029/2010GB003789 culture 184: 115–132. http://dx.doi.org/10.1016/S0044- Ratnarajah, L., S. Nicol, and A. R. Bowie. 2018. Pelagic iron 8486(99)00316-6 recycling in the Southern Ocean: Exploring the contribu- Štrus, J., N. Žnidaršič, P. Mrak, U. Bogataj, and G. Vogt. 2019. tion of marine animals. Front. Mar. Sci. 5: 1–9. http://dx. Structure, function and development of the digestive sys- doi.org/10.3389/fmars.2018.00109 tem in malacostracan crustaceans and adaptation to differ- Rauschenberg, S., and B. S. Twining. 2015. Evaluation of ent lifestyles. Cell Tissue Res. 377: 415–443. http://dx.doi. approaches to estimate biogenic particulate trace metals in org/10.1007/s00441-019-03056-0 the ocean. Mar. Chem. 171: 67–77. http://dx.doi.org/10. Tang, K. W., V. Turk, and H. P. Grossart. 2010. Linkage 1016/j.marchem.2015.01.004 between crustacean zooplankton and aquatic bacteria. Reinfelder, J. R., and N. S. Fisher. 1991. The assimilation of ele- Aquat. Microb. Ecol. 61: 261–277. http://dx.doi.org/10. ments ingested by marine copepods. Science 251: 794–797. 3354/ame Roman, J., and others. 2014. Whales as marine ecosystem 01424 engineers. Front. Ecol. Environ. 12: 377–385. http://dx.doi. Tang, K. W., R. N. Glud, A. Glud, S. Rysgaard, and T. G. org/10.1890/130220 Nielsene. 2011. Copepod guts as biogeochemical hotspots Sampei, M., A. Forest, H. Sasaki, H. Hattori, R. Makabe, M. in the sea: Evidence from microelectrode profiling of Cala- Fukuchi, and L. Fortier. 2009. Attenuation of the vertical nus spp. Limnol. Oceanogr. 56: 666–672. http://dx.doi.org/ flux of copepod fecal pellets under Arctic Sea ice: Evidence 10.4319/lo.2011.56.2.0666 for an active detrital food web in winter. Polar Biol. 32: Thodesen, J., T. Storebakken, K. D. Shearer, M. Rye, B. 225–232. http://dx.doi.org/10.1007/s00300-008-0523-z Bjerkeng, and B. Gjerde. 2001. Genetic variation in mineral Sarthou, G., D. Vincent, U. Christaki, I. Obernosterer, K. R. absorption of large Atlantic salmon (Salmo salar) reared in Timmermans, and C. P. D. Brussaard. 2008. The fate of bio- seawater. Aquaculture 194: 263–271. http://dx.doi.org/10. genic iron during a phytoplankton bloom induced by natu- 1016/S0044-8486(00)00525-1 ral fertilisation: Impact of copepod grazing. Deep-Sea Res. Tortell, P. D., M. T. Maldonado, J. Granger, and N. M. Price. Part II Top. Stud. Oceanogr. 55: 734–751. http://dx.doi.org/ 1999. Marine bacteria and biogeochemical cycling of iron 10.1016/j.dsr2.2007.12.033 in the oceans. FEMS Microbiol. Ecol. 29: 1–11. http://dx. Schmidt, K., and others. 2011. Seabed foraging by antarctic krill: doi.org/10.1016/S0168-6496(98)00113-5 Implications for stock assessment, bentho-pelagic coupling, Turner, J. T. 2015. Zooplankton fecal pellets, marine snow, phyto- and the vertical transfer of iron. Limnol. Oceanogr. 56: detritus and the ocean’s biological pump. Prog. Oceanogr. 1411–1428. http://dx.doi.org/10.4319/lo.2011.56.4.1411 130: 205–248. http://dx.doi.org/10.1016/j.pocean.2014.08.005 212 Le Mézo and Galbraith The fecal iron pump Twining, B. S., and N. S. Fisher. 2004. Trophic transfer of trace Wotton, R. S., and B. Malmqvist. 2001. Feces in aquatic eco- metals from protozoa to mesozooplankton. Limnol. Ocean- systems. Bioscience 51: 537–544. http://dx.doi.org/10. ogr. 49: 28–39. http://dx.doi.org/10.4319/lo.2004.49.1.0028 1641/0006-3568(2001)051[0537:FIAE]2.0.CO;2 Twining, B. S., and others. 2014. Differential remineralization of major and trace elements in sinking diatoms. Limnol. Conflict of Interest Oceanogr. 59: 689–704. http://dx.doi.org/10.4319/lo.2014. None declared. 59.3.0689 Submitted 20 September 2019 Wang, J., and W.-X. Wang. 2016. Novel insights into iron reg- ulation and requirement in marine medaka Oryzias Revised 16 March 2020 melastigma. Sci. Rep. 6: 1–11. http://dx.doi.org/10.1038/ Accepted 20 August 2020 srep26615 Associate editor: Maria Maldonado 213