On the move: sloths and their epibionts as model mobile ecosystems

ABSTRACT Sloths are unusual mobile ecosystems, containing a high diversity of epibionts living and growing in their fur as they climb slowly through the canopies of tropical forests. These epibionts include poorly studied algae, arthropods, fungi, and bacteria, making sloths likely reservoirs of unexplored biodiversity. This review aims to identify gaps and eliminate misconceptions in our knowledge of sloths and their epibionts, and to identify key questions to stimulate future research into the functions and roles of sloths within a broader ecological and evolutionary context. This review also seeks to position the sloth fur ecosystem as a model for addressing fundamental questions in metacommunity and movement ecology. The conceptual and evidence‐based foundation of this review aims to serve as a guide for future hypothesis‐driven research into sloths, their microbiota, sloth health and conservation, and the coevolution of symbioses in general.


I. INTRODUCTION
Sloths spend much of their lives hanging from trees in Central and South America and are unique in that they have some of the slowest metabolisms of all mammals . There are six extant species of sloths in two genera: two-fingered (Choloepus spp., family Choloepodidae) and three-fingered (Bradypus spp., family Bradypodidae) (Slater et al., 2016). Historically, the names 'two-toed' and 'three-toed' have been used, although this is a misnomer; we use 'two-fingered' and 'three-fingered' herein because all sloths have three toes but differ in the number of 'fingers' they have on their upper limbs. Despite both genera being slow-moving arboreal folivores, two-and three-fingered sloths are very different as revealed by molecular, morphological, and behavioural data (Table 1). Mitogenome and ancient collagen DNA phylogenetic analyses have revealed that these two sloth genera diverged 31 million years ago ( Fig. 1; Delsuc et al., 2019;Presslee et al., 2019). Both sloth genera host an array of largely unexplored symbioses (i.e. persistent, physical associations; Bronstein, 2015) involving taxonomically diverse microorganisms and arthropods in a multi-trophic assemblage within their fur or pelage (Aiello, 1985;Gilmore, da Costa & Duarte, 2001;Suutari et al., 2010;Higginbotham et al., 2014). The structure of sloth hair is also unusual, being characterized by cracks or grooves that are hypothesized to facilitate algal growth (Aiello, 1985;Suutari et al., 2010), giving them a distinct green coloration in the wild. With a complex and largely self-contained community of interacting epibionts defined by a boundary layer of fur, the sloth holobiont (host + associated biota) likely functions as an ecological unit in space (Jax, 2006). As sloths move slowly through the greater forest ecosystem, they can be considered 'mobile ecosystems' or dynamic, moving islands of biodiversity (cf. Heaney, Balete & Rickart, 2013;Borregaard et al., 2017). As such, they could represent unique systems to investigate questions in host-epibiont/host-microbiome ecology and coevolution within an unusual spatiotemporal/ movement regime not typical of sessile organisms or fastmoving animals.
We aim to highlight how studying sloths and their epibionts may be useful in addressing fundamental questions in microbial and metacommunity ecology, movement ecology, microbiome science, and the evolution of symbioses. We summarize what is known about the basic biology of sloths as it relates to their epibionts, and review evidence (or lack thereof) in support of several speculative conclusions that have accrued in the literature that have unfortunately led to misconceptions now canonized in the popular media (Meier, 2013;Greenwood, 2014;Lewis, 2014;Woollaston, 2014). We aim to challenge speculations that lack clear empirical support, articulate gaps in our understanding of the sloth as a mobile ecosystem (focused particularly on sloth fur as an ecosystem), and make suggestions for future directions in sloth research.

II. THE SLOTH AS A MODEL MOBILE ECOSYSTEM
All animals possess an assemblage of other species that live on or within them, the majority of which are microbial. When found within (as with gut microbiomes) these species often have a profound influence on host biology (McFall-Ngai, 2015; Barko et al., 2017). As for other mammals, the gut microbiome of sloths is believed to play an important role in sloth health and be influenced by diet (Delsuc et al., 2014;Dill-McFarland et al., 2016). However, it is the rich diversity of epibiotic symbionts on sloth fur that is most distinctive about the sloth holobiont and the focus of this review. Unlike the gut microbiome, which is shielded from the environment except through host-driven dietary intake, the sloth fur ecosystem is open to the larger forest ecosystem through which the sloth moves. In addition to microbes, a variety of arthropods are an integral part of the fur multi-trophic community (discussed in Section III). Similar to the pitcher plant (Boynton, 2012;Bittleston et al., 2018;Miller, Bradshaw & Holzapfel, 2018b), which contains an elaborate food web of predators, prey, and detritivores that reside within a leafy 'cup' and is an entire ecosystem unto itself, sloth fur is a relatively self-contained system of taxonomic diversity and trophic levels.
(1) Unique sloth traits Two-and three-fingered sloths have many similar traits (Table 1), some that are hypothesized to have evolved   convergently. Both groups of sloths have evolved suspensory posture, long, sharp claws for gripping branches and for territorial fights, and a modified skeletal structure to suit their slow, arboreal lifestyle (Miller, 1935;Montgomery & Sunquist, 1978;Mendel, 1981Mendel, , 1985Nyakatura & Fischer, 2011;Nyakatura, 2012;Pauli et al., 2016;Olson et al., 2018). Suspensory posture, and the many anatomical adaptations that have arisen for efficient suspensory locomotion in trees (see Supplementary Videos S3 and S4), are the most clearly convergent traits, given that no known fossil sloths were considered suspensory (Nyakatura, 2012). Evidence for a convergently evolved slow metabolism and diet is lacking, however. While it is not clear if ground sloths had cracked/grooved hair, this is a distinctive trait of all extant sloth species and has not been found on the closest relatives of sloths, armadillos and anteaters (Aiello, 1985; discussed further in Section III.1a), nor for any other mammal. The only other known mammals with epibiotic algal growth are polar bears in zoos (Lewin & Robinson, 1979) and manatees (Bledsoe et al., 2006), although they do not appear to have hair with crevices.
(2) Geographical range, movement, and behaviour Geographically, sloths are found throughout the neotropical forests of Central and South America, occupying a native range from Guatemala south through Peru and Brazil (Fig. 2). Although it is commonly thought that sloths are quite sedentary, they have been observed to move regularly throughout the forest canopy at rates of up to 0.5 km/h in short bursts (Sunquist & Montgomery, 1973). In one study based on radio-telemetry observations, 54% of C. hoffmanni sloths were observed to move ≥38 m between daily locations (Sunquist & Montgomery, 1973) and <10% of observed sloths were found on the same tree on successive days. These two-fingered sloths are nocturnal with the majority of their activity occurring in bouts lasting 2-6 h, with total activity averaging 7.6 h per day (Table 1). In a cacao agroecosystem, C. hoffmanni was found in 101 different tree species and eating from 34 species (Vaughan et al., 2007). By contrast, B. variegatus sloths tended to travel shorter distances, with 89% observed to travel <38 m per day (Sunquist & Montgomery, 1973). Unlike C. hoffmanni, these three-fingered sloths are active throughout the day and night and for a greater proportion of the day (10 h) but are four times more likely to remain on the same focal tree on successive days (Sunquist & Montgomery, 1973). This is consistent with observations of B. variegatus being found in and eating from fewer tree species than C. hoffmanni (71 and 15, respectively), and their preference to return more frequently to feed on specific trees, especially Cecropia spp. (Vaughan et al., 2007;Neam & Lacher, 2015;Garcés-Restrepo, Peery & Pauli, 2019b). B. variegatus is more likely to have bouts of activity of shorter duration than C. hoffmanni, with the majority being <1 h (Table 1). Like C. hoffmanni, however, bouts of activity lasting 2-6 h each account for the majority of the total time active during the day.
Home range sizes for sloths varies depending on species and sex, estimates of which are based largely on observations of C. hoffmanni and B. variegatus in mixed-use landscapes centered around cacao plantations (see Table 1). Median home ranges for C. hoffmanni and B. variegatus sloths are estimated to be 4-7.5 and 5 ha, respectively, which are generally within estimates predicted by a body size-space use scaling law for mammalian herbivores (Jetz et al., 2004; Table 1). However, individual home range observations are extremely variable, with a long-tail distribution (and mode <2 ha) (Sunquist & Montgomery, 1973;Vaughan et al., 2007;; these data were typically acquired over a 1-year period and it is possible that range estimates could increase (and variability decrease) with longer observation times (cf. Vaughan et al., 2007). Nonetheless, males appear to have larger home ranges than females in general.
Based on what we know about C. hoffmanni and B. variegatus, sloths move across the forest not just laterally, but up and down the canopy column. Sloths favour floristically diverse and structurally complex forest structures (vertically and horizontally), with dense canopy cover and variable tree heights (Montgomery & Sunquist, 1978;Mendoza et al., 2015;Neam & Lacher, 2015. As heterotherms, they prefer to rest and feed high up in the forest canopy (Neam & Lacher, 2015), moving in and out of canopy shade/sunlight and up and down trees to regulate their body temperatures as needed depending on ambient temperatures throughout the day (Montgomery & Sunquist, 1978;Pauli et al., 2016).
Sloths also exhibit the unusual behaviour of descending all the way down to the ground to defecate about once a week (Montgomery & Sunquist, 1975;Waage & Montgomery, 1976;Montgomery & Sunquist, 1978;Voirin et al., 2013). Many ideas have been proposed to explain this unusual defecation behaviour. It has been proposed that defecating on the ground (as opposed to letting dung drop from the canopy of trees) is a strategy that sloths use to remain undetected, since being quiet and hidden seems to be their predominant life strategy and defecating from the canopies of trees presumably may cause a disturbance that attracts predators (S. Trull, unpublished data). However, there is no evidence that descending to the base of the tree is risky to the sloth, especially since the majority of their predators, harpy eagles (Harpia harpyja), spectacled owls (Pulsatrix perspicillata), ocelots (Leopardus pardalis), and tayra (Eira barbara), can also detect and attack them from the tree canopy, often by knocking them to the ground where they proceed to eat them (Beebe, 1926;Izor, 1985;Bezerra et al., 2009;Voirin et al., 2009). Other theories include proposed benefits from fertilizing their most frequently used trees, communicating with other sloths through social latrines, trying to hide their scent from predators, sustaining a three-way mutualism with moths and algae (discussed further in Section III.2c), or deriving nutritional benefits from consuming soil while on the ground (Beebe, 1926;Krieg, 1939;Goffart, 1971;Voirin et al., 2013). However, observational data suggest that threefingered sloths do not frequently eat soil (S. Trull, unpublished data), and no data exist in support of the other theories.

III. COMPONENTS OF THE MOBILE ECOSYSTEM
(1) Algae The green hue of sloths arises from green algae that grow on sloth hair (Aiello, 1985;Suutari et al., 2010). Cyanobacteria may also contribute to this greenish hue, although only one species, Oscillatoria pilicola, has been identified to the species level to date (Table 2; Wujek & Lincoln, 1988). DNA sequences for red algae have also been found on sloths (Table 2; Suutari et al., 2010). For this review, we use the term 'algae' to refer broadly to eukaryotic algae and cyanobacteria unless specifically distinguished. It is not clear if algae are resident on all sloths in the wild. One study found that 73% of the 74 sampled sloths had visible algae on their fur identified via eye or microscope [Bradypus variegatus (N = 18), Bradypus tridactylus (N = 12), Bradypus pygmaeus (N = 12), Bradypus torquatus (N = 8), Choloepus hoffmanni (N = 22), Choloepus didactylus (N = 2)] (Suutari et al., 2010). However, neither sloth age, season of sampling, nor location were accounted for, and included in this analysis were captive sloths from zoos, which may lack native epibionts (likely due to being bred in captivity, bathed, or being kept in an enclosed habitat away from potential microbial symbionts in their native habitat). Given the limited and uneven sampling of this study, this value of 73% should not be interpreted as a definitive statistic. It is also generally overlooked that 'brown' sloths may actually host epibiotic algae even though not visibly green to the naked eye (Goffart, 1971): such algae may simply be in a dormant or non-green state when moisture is limited. In fact, wetting of 'brown' sloth hair results in a rapid greening within seconds to minutes (Fig. 3), akin to what is observed with the wetting of desiccated biological soil crusts (Abed et al., 2014;Pietrasiak, 2014).   (a) Sloth hair structure and algal growth The morphology of sloth hair has the potential to influence the extent and composition of symbiotic growth. Threefingered sloth hair has transverse cracks that increase in quantity and depth as sloths age ( Fig. 4; Aiello, 1985;Wujek & Cocuzza, 1986). The hairs swell considerably when wet, and it has been hypothesized that moisture that is retained within cracks sustains algal growth on the surface of the hairs (Aiello, 1985). It has also been hypothesized that this hair-based absorption of water may help buffer changes in sloth body temperature . It does not appear that the algae grow directly within the cracks, which would potentially limit access to photosynthetic radiation, but rather grow on the smooth outer surface of the hair (Aiello, 1985). It remains unknown whether algae directly colonize hair strands with pre-existing cracks and/or if they contribute to hair crack development. By contrast, twofingered sloth hair has vertical grooves and does not appear to absorb as much water; algae appear only to be found within the grooves instead of coating the entire hair ( Fig. 4B; Aiello, 1985;Wujek & Cocuzza, 1986). Differences in hair architecture may be responsible for the observed differences in fur amplicon sequencing surveys between the two genera of sloths (Aiello, 1985;Suutari et al., 2010). Although increased absorptive properties due to unusual hair structure are not limited to sloths (Kingdon et al., 2012), the unique cracked/grooved hair structure of sloths seems to facilitate symbiotic algal growth unlike any other mammal (Aiello, 1985). It is unknown whether algal and fungal species typically found on sloth hair are able to grow on texturally smooth hair. Whether such hair cracks/grooves co-evolved with the associated microbes or is a convergently evolved trait among two-and three-fingered sloths remains an open question.

(b) Identification of sloth algae
Morphological identification of sloth algae has yielded confusing results; in most cases, the sloth species from which algae have been derived was not recorded (Table 2). In a conference abstract by Thompson (1972), many sloth-furassociated algae and cyanobacteria were listed, identified via morphology. However, algal and cyanobacterial species can be very hard to distinguish morphologically; DNA-and polyphasic-based methods are typically required to make clear taxonomic assignments (Leliaert et al., 2014;Wilmotte et al., 2017). Unfortunately, no follow-up confirmations of Thompson's (1972) identifications exist in the literature and Thompson did not specify from which sloth species these algae were obtained. Thompson identified two species of Oscillatoria and one of Nostoc, but it is not clear if either of these Oscillatoria are the same as the Oscillatoria pilicola identified and described by Wujek & Lincoln (1988) on both the fur of B. variegatus and C. hoffmanni. The genus Fischerella, three coccoid green algae (including Dictyococcus bradypodis and Chlorococcum choloepodis), three species of Trentepohlia, two of Stichococcus, and one of Nannochloris were identified (Table 2; Thompson, 1972;Wujek & Timpano, 1986). Rufusia, a red alga named by Wujek & Timpano (1986), was also identified on both B. variegatus and C. hoffmanni. Trichophilus welckeri, the best-known sloth green alga, was first identified on sloths in 1887 (Weber-van Bosse, 1887). Trichophilus is in the class Ulvophyceae and is characterized by small (3-13 μm) thick-walled cells with numerous, small, discoid chloroplasts that lack pyrenoids ( Fig. 5; Table 2; Printz, 1964;Suutari et al., 2010). While solely morphologybased taxonomic identifications must be taken cautiously, many of the aforementioned findings seem to be unknown to most modern readers of the sloth literature. Nonetheless, it appears that the diversity of green algae and cyanobacteria may be far greater than is suggested by more recent studies that focus on T. welckeri and its role in the sloth hair ecosystem (Pauli et al., 2014). Other species of algae should be taken into consideration, however, to understand how the community of photobionts is functioning and impacting its accompanying fungal and bacterial epibionts, arthropods, and the sloth itself.
Amplicon metagenomic studies of sloth fur to date reveal a variable array of algae across and within different sloth species. Using amplicon-based techniques, T. welckeri was found on the fur of B. variegatus, the pale-throated sloth, Bradypus tridactylus, and the pygmy three-fingered sloth, Bradypus pygmaeus (Suutari et al., 2010). No other green algal species were identified on these sloths from this 18S amplicon sequencing study (Suutari et al., 2010; Table 3) despite the prior observations by Thompson (1972), Wujek & Timpano (1986) and Wujek & Lincoln (1988). This discrepancy may be due in part to the use of captive sloths from zoos that potentially lack native epibionts (as mentioned earlier) and/or the intrinsic limitations of the use of 18S ribosomal DNA (rDNA) as a barcode for resolving algal taxa (Hall et al., 2010). T. welckeri has not yet been found environmentally (Suutari et al., 2010), although this may be a consequence of insufficient environmental sampling across the sloths' geographical range and within the canopies of trees. The maned three-fingered sloth, Bradypus torquatus, hosts a variety of algae belonging to genera known to be terrestrial, e.g. Trentepohlia and Myrmecia (Table 3; Suutari et al., 2010). C. hoffmanni, and B. tridactylus host Trichophilus spp. as well as terrestrial green algae from their surroundings (Table 3; Suutari et al., 2010).
The 18S sequences for Trichophilus spp. found in association with B. variegatus, B. pygmaeus and B. tridactylus were found to cluster separately from Trichophilus sequences obtained from C. hoffmanni (Suutari et al., 2010). Trichophilus spp. from Bradypus and Choloepus differ in cell size, and B. variegatus and T. welckeri phylogenies are consistent with codivergence, which has led some to propose that B. variegatus and T. welckeri have coevolved (Suutari et al., 2010;Fountain et al., 2017). However, matching phylogenies is an insufficient demonstration of reciprocal coevolution (Janzen, 1980;Anderson, 2015). The differences in hair structure discussed earlier may impact differential colonization of sloth hair and the poorly characterized biogeography of environmental sources of sloth algae might explain the underlying phylogenetic concordance.
(c) Algal benefits Several hypotheses have been proposed for how algae might benefit sloths, however, they all lack concrete empirical support, and in fact, it is not clear if algae provide any benefit to sloths. It is possible that it is simply a commensal relationship, and sloths may have so much algae in their fur because they do not have the means to clean themselves. Despite this, it is widely believed that fur algae provide a camouflage benefit to the sloth (Aiello, 1985;Suutari et al., 2010;Pauli et al., 2014), but no studies have been pursued to test this hypothesis. As discussed above, sloth fur coloration can change: they are primarily green during the rainy season when their hair is regularly wet (Fig. 6A), and in the dry season, many sloths lose their greenish hue and appear brown or grey (Britton, 1941;Gilmore et al., 2001). Direct observations of brown/grey sloths in their native canopy suggest that they are very well camouflaged with this colour scheme, blending in with the branches, trunks, and dead leaves of trees (Fig. 6C, D), as well as resembling ant and termite nests ( Fig. 6B; Goffart, 1971).
Fur algae have also been proposed to serve as a nutritional food source for sloths based on remnants of green algal cells being found in their stomach contents (Pauli et al., 2014). However, there is little evidence to support the idea that sloths 'farm' algae on their fur that they then consume, or that they consume sufficient quantities of fur algae to be nutritionally beneficial (discussed further in Section III.2c). Other proposed hypotheses in the literature for how algae could benefit sloths include: (i) algae being a source of thermal insulation (Britton & Atkinson, 1938;Goffart, 1971;Montgomery & Sunquist, 1978;Aiello, 1985); (ii) algae providing some yet unidentified chemical benefit to overall sloth health (Aiello, 1985); (iii) algae producing exopolymeric substances to facilitate beneficial bacterial growth (Suutari et al., 2010); and (iv) algae acting as a sunscreen (Suutari et al., 2010). T. welckeri has been found to produce an UVabsorbing mycosporine-like amino acid believed to protect algae from UV radiation (Karsten et al., 2005); whether this protection extends to sloths with T. welckeri on their fur (beyond the intrinsic UV-protection of fur itself) is unknown. To our knowledge, none of these hypotheses are supported by any evidence-based rationale.
(2) Arthropods (a) Biting arthropods Sloths are also host to a wide range of arthropods living in their fur including parasitic, bloodsucking and biting arthropods such as mosquitoes and sandflies, triatomine bugs, lice, mites, and ticks (Gilmore et al., 2001). Six species of ticks have been found on two-and three-fingered sloths, all from the genus Ambylomma, but only two species, Ambylomma geayi and Ambylomma varium, appear specialized for living on sloths as these ticks are rarely found on other hosts (Waage & Best, 1985). Tick infestation can be extremely high. At the Instituto Nacional de Pesquisas da Amazonia in Manaus   Fig 7. The sloth-associated scarab beetle "Uroxys gorgon Arrow, 1933." (A) Collected live from the fur of a Bradypus variegatus (brown-throated three-fingered sloth), and (B) a mounted specimen (Larsen, 2015), used with permission under Creative Commons License CC BY-NC 3.0. Table 3. Sloth species and associated algal epibionts identified to date. Those with an asterisk following the genus have thus far only been found on sloths and not yet on other environmental substrates. Data are from Suutari et al. (2010; clarified in some cases through personal correspondence with M. Suutari and J. Blomster). Cyanobacteria are indicated by superscript C . Eleven genera not listed in the table, Chlorococcum, Collinsiella, Dictyococcus, Fischerella C , Nannochloris, Nostoc C , Planophila, Pseudendoclonium, Stichococcus, Trichosarcina, and Ulothrix, were found on sloths, but are of an unidentified origin (Thompson, 1972;Wujek & Timpano, 1986). Note that Myrmecia below is a genus of green algae associated with lichens, not the genus of ants On sloths and their epibionts (Brazil), 99% of three-fingered and 86.7% of two-fingered sloths carried Ambylomma spp. (Waage & Best, 1985). Nothing is known about how A. geayi or A. varium find a host sloth and no correlation has been found between the numbers of ticks at any life stage on a sloth or seasonal differences in rainfall (Gilmore et al., 2001). The blood-sucking mites, Liponissus inheringi, Lobalges trouessarti, and Edentalges bradypus have been identified on three-fingered sloths (Waage & Best, 1985) and the mite Edentalges choloepi has been found on Linnaeus's two-fingered sloth, Choloepus didactylus (Fain, 1964). Whether ectoparasite loads impact the health of a sloth is an open question.

(b) Commensals and beetles
Many commensal arthropods are found in association with these slow-moving mammals. It is quite possible that the algae on sloth fur serves as a food source for these commensal arthropods considering that mites and other insects display algophagy (Mckenna et al., 2015;Seniczak et al., 2016). Cockroaches have been found in sloth fur (Britton, 1941), although this may be quite rare (S. Trull, unpublished data). Adults of several scarab beetle species are frequently found in the fur of three-fingered sloths (of which the beetle in Fig. 7 is an example), but have not been reported to be associated with Choloepus (Ratcliffe, 1980;Gilmore et al., 2001). The scarab beetles occur near the elbow or on the flanks behind the knees, buried deep inside the fur. The beetles found living on sloths are considered commensal because they are phoretic coprophages: the beetle larvae (and possibly adults) feed on sloth dung and they do not appear to harm the sloths (Ratcliffe, 1980;Gilmore et al., 2001). About a thousand such beetles (Trichillum adisi) have been found in the fur of a single brown-throated three-fingered sloth (B. variegatus) collected on Curari Island in the Central Amazon region (Waage & Best, 1985). Beetles of the genus Uroxys have been recorded from sloths in Bolivia, Brazil, Colombia and Panama (Waage & Best, 1985). Despite the ubiquity of beetle-sloth interactions, little is known about the dispersal and density fluctuations of these beetles on sloths, although in Panama, there seem to be higher numbers of beetles during the rainy season (Wolda & Estribi, 1985). It has been suggested that the beetles have dispersal flights at the beginning and end of the rainy season and that part of the population might enter reproductive diapause and disperse from the sloths to sites with some moisture; they presumably resume reproduction at the end of the dry season and return to the sloths (Wolda & Estribi, 1985). Just as there are no data to substantiate an effect of parasite load on sloths, no analysis has been performed to understand the effect of these suspected commensal arthropods or of total arthropod load on sloth health. Likewise, little is known of the potential role these beetles might play in the ecosystem. It is possible that beetles contribute to parasite suppression, secondary seed dispersal, and to nutrient cycling within the sloth fur ecosystem and the larger forest ecosystem (Nichols et al., 2008). It is also possible that some sloth-associated arthropods play a protective and mutualistic role by preying on ectoparasites in sloth fur (cf.

(c) Sloth moths
Sloth moths in the genus Cryptoses have received notable attention as a sloth epibiont. There is appreciable geographic sympatry amongst sloth-associated moth species and several different species may coexist in the fur of a single sloth (Waage & Best, 1985). Various sloth moth species appear to be found on all species of sloths (Bradley, 1982;Waage & Best, 1985;Pauli et al., 2014). Cryptoses choloepi seems to be the most common moth found on B. variegatus and has been studied almost exclusively in relation to this sloth species ( Fig. 8; Supplementary Videos S2 and S3). Female C. choloepi moths that live in B. variegatus fur have been observed to oviposit in the dung of the sloth as the sloth descends to the forest floor to defecate. Moth larvae in early stages spin silken threads between two or three pellets of dung, forming net-like structures from which they feed (Waage & Montgomery, 1976). Upon maturation after 3-4 weeks, newly emerged moths presumably fly from the dung pile into the forest canopy to find a new sloth host (Waage & Montgomery, 1976). In addition to nutritional benefits that sloth moth larvae might receive from feeding on sloth dung, adult moths may consume sloth/algal secretions or microbes on sloth hair (Fig. 8); the sloth moth gut microbiome has yet to be explored but may provide evidence for this. Adult moths are believed to receive a transportation benefit as well as a protection benefit from living in sloth fur (Waage & Montgomery, 1976;Wolda, 1985). The amount of protection moths receive in association with sloths is questionable, however, since brown jays (Psilorhinus morio) have been observed to consume insects off sloth fur (Neam, 2015).
Based on studies to date, it would appear that sloth moths have a commensal relationship with their sloth hosts. However, a three-way mutualism has been proposed involving B. variegatus, their moths, and fur algae, specifically T. welckeri. According to this hypothesis, moths are portals for nutrients, increasing nitrogen levels in sloth fur through defecation, which is believed to promote algal growth (Pauli et al., 2014). T. welckeri-like algae have been found (microscopically) in sloth stomach contents, which has led to the hypothesis that sloths consume these algae to augment their limited diet (see Section III.1c). With this set of observations, the proposal is that sloths are involved in an evolutionary trade-off in which they risk their lives, descending to the ground to defecate, in order to preserve this sloth-mothalgae tripartite mutualism (Pauli et al., 2014). There are at least five potential problems with this hypothesis. First, the method of identifying the alga in this study as T. welckeri by morphology alone is not sufficiently rigorous, especially given that this taxon is often morphologically cryptic and understudied in general (Dudgeon et al., 2017). Second, while the main groups of bacteria that inhabit the gut microbiome of B. variegatus have been identified (Dill-McFarland et al., 2016), no metagenomic study to date has been performed to characterize or confirm the eukaryotic/algal diversity in this species' gastrointestinal tract. Sloths have been observed to lick and eat material off of branches and tree trunks, which may include lichens (Tirler, 1966;S. Trull, unpublished data), thus the algae observed in the stomachs of sloths may have derived from environmental substrates rather than from their fur. Third, thousands of hours of sloth behavioural research recorded during the day and night do not support the idea that sloths lick themselves (like cats) or eat epibiotic algae from their fur (Tirler, 1966;S. Trull, unpublished data). Fourth, only two B. variegatus individuals out of 12 sampled in one location in Costa Rica were identified as having Trichophilus spp. in their stomachs (Pauli et al., 2014). And lastly, if sloth tree-descent and ground-defecation is driven by a need to benefit moths via dung oviposition, one would expect there to be reciprocal fitness benefits provided to the sloth by the moths in order for this behaviour to have evolved or be maintained (Voirin et al., 2013); however, the implied and indirect benefits that sloths might obtain from moth-influenced fur algal growth lack empirical support and may be quantitatively modest.

(3) Fungi
Fungi are known to be associated with sloth hair, but the roles they play in the community ecology of the sloth pelage and in the health of the sloth remain unexplored. A diverse group of Ascomycota and one Basidiomycete (Sporobolomyces subbrunneus) have been identified growing on sloth fur through sequencing and culture-based methods (Suutari et al., 2010;Higginbotham et al., 2014). Only two species of fungi that have been found on sloths have also been found on the bark of trees in sloth habitats (Devriesia staurophora and Mycosphaerella pini; Suutari et al., 2010), although these results are from very limited sampling. These sloth-associated fungi have been found in soil and plants (Arnold & Lutzoni, 2007;   On sloths and their epibionts et al., 2011), so it is possible that sloths are exposed to these fungi when they defecate on the ground or as they eat and interact with leaves and bark (Higginbotham et al., 2014). Nearly 35% of fungal isolates obtained from B. variegatus fur are identical to endophyte strains obtained from plants in the same region (Higginbotham et al., 2014). Given the taxonomic similarity between endolichenic and endophytic plant fungi in the same environments (U'ren et al., 2012), it is plausible that some sloth hair fungi may associate directly with green algae (Higginbotham et al., 2014). Previous studies support that fungi, and these taxa in particular, have intrinsic affinities for associating and forming mutualisms with algae, as seen in lichens [typically slow growing and commonly found on undisturbed trees (Hawksworth, 1988;Arnold et al., 2009)], and other similar symbioses (Hawksworth, 2000;Gareth Jones, Pang & Stanley, 2012;Hom & Murray, 2014;Du et al., 2019).
However, whether sloth-associated fungi are commensals, parasitic, or mutualistic is not clear. Hair-associated fungi from B. variegatus have been shown to display a broad range of inhibitory activities against parasites that cause malaria (Plasmodium falciparum) and Chagas disease (Trypanosoma cruzi), human breast cancer cells, and bacteria, particularly Gramnegative bacteria (Higginbotham et al., 2014). The inhibitory activities of these fungi may thus provide benefits to the sloth as well. Some sloths have clear black fungal growth on their hair (Fig. 9A), which could potentially harm the sloth or outcompete other microbes in the sloth hair ecosystem. Others develop severe fungal infections on their skin that can be detrimental because the infections produce scabs, which then fall off, leaving bare skin that is susceptible to ectoparasites like ticks and mosquitos ( Fig. 9C; Xavier et al., 2008); anecdotally, fungal infections generally correlate with sick sloths (S. Trull, unpublished data). It is unclear whether these infections derive from fungi that are already part of the sloth fur microbiome and/or whether they are a primary cause for sloths getting sick.

(4) Other symbionts
In addition to the algae, arthropods, and fungi that live and thrive within the pelage of sloths, other putative furassociated organisms have been identified through 18S amplicon sequencing; these include euglenozoans, amoebozoans, cercozoans, apicomplexans, dinoflagellates, and ciliates (Table 4; Suutari et al., 2010). To date, nothing is known about the role of these organisms within the sloth hair ecosystem. Apart from the sloth fur cyanobacteria mentioned above (Table 2), fur-associated prokaryotes have not been well documented or sufficiently taxonomically resolved. Surprisingly, a 16S survey of the bacterial diversity on sloth fur has not yet been performed; it will be important to survey the prokaryotes present in the sloth fur ecosystem and to understand the inter-kingdom interactions they may have with the sloth and other fur epibionts. Bacterial epibionts may influence the function of sloth-associated fungi and algae, as they do for fungal endophytes associated with plants (Partida-Martinez & Hertweck, 2005;Hoffman & Arnold, 2010) and lichens (Grube & Berg, 2009;Bates et al., 2011).
Sloths are also carriers of a variety of arthropod-associated viruses ('arboviruses'; e.g. phleboviruses, flaviviruses, encephalitis viruses, and orthobunyaviruses) (Seymour, Peralta & Montgomery, 1983;Gilmore et al., 2001;Medlin et al., 2016;de Oliveira Filho et al., 2020) and insect-borne protozoans (e.g. trypanosomes, such as Leishmania) (Shaw, 1985;Gilmore et al., 2001;Muñoz-García et al., 2019). These pathogens may be transmitted by (freeliving or epibiotic) arthropod bites to blood but may also be found on the skin or within sloth fur. Phlebotomine sandflies that reside in the fur of sloths are known carriers of Leishmania, which causes leishmaniasis in humans (Arias & Freitas, 1978;Herrer & Christensen, 1980;Christensen et al., 1982). C. hoffmanni sloths likely become infected by these trypanosomes in their first few months of life and remain infected for a long time, but appear asymptomatic and do not show signs of pathology (Herrer & Christensen, 1980).

IV. FUTURE DIRECTIONS (1) Transmission of fur epibionts and coevolution
How are sloth fur epibionts acquired and transmitted? There are few to no data on the modes of transmission of the different sloth fur epibionts, or whether there are taxon-specific differences in the degree of vertical versus horizontal transmission.
To determine the extent of epibiont coevolution with sloth hosts, however, it is critical to determine the degree of vertical transmission of the epibiont community and reciprocal adaptation (Bordenstein & Theis, 2015;Meng et al., 2018;Roughgarden et al., 2018;Simon et al., 2019). Sloth algae, the most studied fur epibiont, is believed to be transmitted vertically from mother to baby (Beebe, 1926;Britton, 1941;Suutari et al., 2010) although this has not been tested directly. Horizontal transmission from environmental species pools to the sloth cannot be ruled out given the absence of data characterizing these pools. This is true for all fur epibionts that have been hypothesized to be specific to sloths. A mixed mode of epibiont transmission is likely, however, given that vertical and horizontal modes represent extreme cases . Obligate symbionts generally rely on vertical transmission , although there are also no compelling data on whether the algae on sloths are obligately or facultatively associated. It is tempting to imagine that the peculiar cracks/grooves of sloth hair may be a coevolved adaptation of the sloth with fur algae, but it is possible this may simply be a trait exapted by algae, given that cracked hair retains water that may aid sloths in thermoregulation (see Section III.1a).
The fact that over a third of sampled fungi from sloth fur were found to be identical to plant endophytic fungi from Biological Reviews 96 (2021)  foliage in the same regional habitat and that other identified fungal genera can also be found in soil and with plants (Section III.3) suggests that fur microbes could in general be acquired from the environment. This may be particularly true for arthropods that are more mobile than microbes. These results highlight the importance of carefully characterizing environmental species pools to establish a null model that can be used to test hypotheses about vertical transmission or coevolution (cf. Fountain et al., 2017), efforts that have been lacking for sloth algal surveys to date. Although several arthropods appear to be specific to sloths, including the ticks A. geayi and A. varium (Waage & Best, 1985), the scarab beetle U. gorgon (Fig. 7), and sloth moth C. choloepi (Waage & Montgomery, 1976;Fig. 8), their transmission across generations remains unknown. The notion that sloth moths could have co-evolved with B. variegatus and the green alga T. welckeri would require a high degree of partner fidelity (Archetti et al., 2011;Kaltenpoth et al., 2014). However, it is uncertain how after 3-4 weeks of maturation at a site of sloth defecation (Waage & Montgomery, 1976; Section III.2c) the moths emerge to find the original sloth host (rather than a different sloth) given the extent of sloth movement during this time (Section II.2).
In theory, horizontal transmission of epibionts could also occur between different sloths. However, sloths are considered solitary animals (Taube et al., 1999;Soares & Carneiro, 2002;S. Trull, unpublished data) and they generally do not interact with animals of other species, except for the occasional bird eating an insect off the sloth (Neam, 2015). On rare occasions, individuals from both two-and three-fingered sloths have been observed to share the same or adjacent trees (Silva et al., 2013), including one extreme case of up to five B. variegatus sloths on a cacao farm feasting on the same Cecropia tree for over a month until the tree was completely defoliated (Vaughan et al., 2007), although this latter example may be far from representative of behaviour in the natural habitat. There is one brief and aggressive encounter documented in the literature between two males possibly related to defending mating territory (Greene, 1989), but this is consistent with sloths' general preference to live alone. It is thus unlikely that sloth epibionts are typically transmitted through social contact with other sloths or other animals.
Sloths of the same species do, however, interact during two phases of sloth life history at which time fur epibionts could be transmitted: mating and early development. Sloths mate with the male on the back of the female or face-to-face, and can copulate for up to 7 min (Bezerra et al., 2008;Dias et al., 2009;Richard-Hansen & Taube, 1997;S. Trull, unpublished data). Close physical contact during copulation could allow for the transmission of epibionts, especially mobile epibionts, such as arthropods, along with any microbes they might carry. Between the birth of young (gestational period of 5-10 months), sloths mate every 10-15 months for a total period of 20 years (Taube et al., 2001); this amounts to approximately 10 matings over the life of a sloth, often with a different partner. The role of sex and the 'reproductive microbiome'i.e. the microbiome associated with the reproductive system of parents that may inoculate offspring (Rowe et al., 2020)on the transmission of fur epibionts between sloths remains to be elucidated.
Sloths give birth to their young in the canopies of trees, and newborn sloths immediately cling to the fur of the mother sloths' abdomen for a continuous period of 5-7 months (Ramirez et al., 2011). Young sloths generally cling to the abdomen of their mother, not her back. However, juvenile sloths do climb onto the back and sides of the mother when she is stationary (Soares & Carneiro, 2002;S. Trull, unpublished data). It is not clear what microbes grow on the abdomen of sloths, since all sloth hair microbiome studies to date have sampled from the greenest parts of the sloth, generally the head, shoulder, or back (Suutari et al., 2010;Pauli et al., 2014). Since juvenile sloths remain on their mother for so many months, fur epibionts are likely vertically transmitted due to protracted close contact. At the very least, mothers dictate the exposure of their young to environmental species pools by the nature of their own movement throughout the forest canopy (Soares & Carneiro, 2002). Frequent sampling and analysis of sloth hair from a mother and her young throughout the period of maternal care and after the juvenile has separated from the mother would be very helpful towards resolving questions about epibiont transmission.
Sloths spend upwards of 70% of their waking hours resting in trees (Chiarello, 1998;Urbani & Bosque, 2007). They are often in direct contact with tree bark and leaves during their sleeping and resting hours, as they can be routinely found lying on branches or reclining against a branch or the trunk of a tree (S. Trull, unpublished data). Thus, transmission of biota from trees to sloths and vice versa is very likely, although there are few formal data to support this hypothesis. The phyllosphere is teeming with microorganisms, such as bacteria, archaea, fungi, and algae (Vacher et al., 2016), and by using metagenomic tools, one could in principle compare the structure and function of microbial communities on sloths and their surrounding canopy environment (Rastogi, Coaker & Leveau, 2013;Baldrian, 2017;Hassani, Dur an & Hacquard, 2018). Sloths also interact with soil when they descend to the base of a tree to defecate once a week (Voirin et al., 2013;Pauli et al., 2014), where they could also acquire or disperse epibionts. The arthropods that reside in sloth fur may also be vectors that transmit microbial epibionts to and from sloths (Laroche, Raoult & Parola, 2019), especially for microbes that may have no apparent environmental species pool in the immediate surroundings of the sloth.
(2) Sloth fur ecosystem benefits It is undeniable that sloth fur is an unusual and rich nexus of biodiversity, but does this fur ecosystem benefit the sloth in any way? Several ideas have been proposed for how algae could benefit the sloth (see Sections III.1c and III.2c); while all are problematic and/or lack empirical support, the notions that algae might provide some camouflage or chemical benefit to the sloth may be plausible. Based on these notions and inspired by the curious visual resemblance of hanging sloths to ant/termite nests (Fig. 6B), we suggest another hypothesis for how fur microbes could benefit the sloth. Azteca ants form a well-known defensive mutualism with Cecropia trees (Janzen, 1969;Berg & Rosselli, 2005;Marting et al., 2018), a genus of trees that B. variegatus frequently use for food and refuge (Fig. 10;Vaughan et al., 2007;Ramirez et al., 2011;Neam & Lacher, 2015;Garcés-Restrepo et al., 2019b). Azteca ants fiercely defend these trees from herbivores (Schupp, 1986). While it is unknown whether these ants are effective at preventing sloths from eating the leaves of the Cecropia (Fig. 10D), anecdotal evidence suggests that sloths are not deterred by these notoriously aggressive biting ants (S. Trull & P. Marting, unpublished data). Given the broad abilities of microbial volatile organic compounds (mVOCs) to deter or modulate insect behaviour (Davis et al., 2013;Engl & Kaltenpoth, 2018), it is possible that semiochemicals produced by sloth hair microbes may act to repel Azteca ants. Such mVOCs may also play a role in scent-based sloth-sloth recognition and mating (see Silva et al., 2013). While mVOCs from plants (Leach et al., 2017), bacteria, and fungi (Dickschat, 2017;Lemfack et al., 2017) have been investigated, the capacity for algae to produce such compounds has been little explored (Achyuthan et al., 2017;Lemfack et al., 2017). Given the prevalence of bacteria (e.g. Streptomyces and Myxobacteria) that produce mVOCs in addition to other diverse compounds (Audrain et al., 2015;Lemfack et al., 2017;Veselova, Plyuta & Khmel, 2019) and unexplored algal and fungal mVOCs, the sloth fur microbiome may be a reservoir for novel mVOCproducing microbes.
Microbes are known to play a fundamental role in the development and immune function of most animals (McFall-Ngai, 2014Bosch, Guillemin & McFall-Ngai, 2019) and this may also be true for sloths. Sloths in captivity face many health challenges (de Stefani Muna o Diniz & Oliveira, 1999); misinformed practices at sloth rehabilitation facilities and zoos, such as bathing sloths without a specific need, could be ridding them of beneficial fur epibionts and disrupting fur ecosystem balance in a manner that negatively impacts sloth well-being (cf. Hooks & O'Malley, 2017;Levy et al., 2017). The fungal infections afflicting wild sloths (Section III.3; Fig. 9) may serve as a practical target for understanding the basis of sloth disease regarding the normal functions and composition of the sloth fur ecosystem. Because symbiotic interactions can be context dependent and mutualists/commensals can potentially become parasites (Bronstein, 1994;Kogel, Franken & Hückelhoven, 2006;Leung & Poulin, 2008;Jones et al., 2015;Akçay, 2017;Vostinar & Ofria, 2019), it is possible that these fungal infections stem from resident or dormant members that become pathogenic due to environmental shifts or fur ecosystem imbalance. It is also possible that the sloth fur ecosystem endows sloths with a resilience against external pathogens.
(3) Access to a unique ecological regime in time and space It is becoming evident that explicit consideration of spatial, temporal, and phylogenetic scales (specifically ideas of granularity and extent) along with system nestedness will be critical to elucidating both the patterns and mechanisms of community assembly in ecosystems in which microbes play a foundational role (Addicott et al., 1987;Wiens, 1989;Wang & Loreau, 2014;Shade et al., 2018;Ladau & Eloe-Fadrosh, 2019). For sloths that are effectively foci of epibiont biodiversity, it may be fruitful to consider how much of the sloth fur community could be understood from the perspective of island biogeography (MacArthur & Wilson, 1967;Bell et al., 2005;Peay et al., 2007;Wilson, 2010;Belisle, Peay & Fukami, 2012;Glassman et al., 2017;Proctor & Relman, 2017). Sloths may be a good model for examining interactions at different hierarchical levels whereby biotic feedbacks between microbes and higher trophic levels of an ecosystem are explicitly considered in understanding how host ecosystems are shaped and structured (Leibold et al., 2004;Carmona et al., 2016;Seibold et al., 2018;Miller, Svanbäck & Bohannan, 2018a;Leibold & Chase, 2019;Liu et al., 2019).
We have referred to the sloth as a 'mobile ecosystem' to highlight the fact that sloths experience life and movement within an unusual regime of time and space, unlike most other macro-organisms. They may serve as genetic and biotic 'mobile links' within neotropical forests (Lundberg & Moberg, 2003). The slow movements of sloths through their geographical range and the vertical column of the forest canopy may allow us to examine an ecological and spatiotemporal regime not typically experienced by sessile (e.g. plants/ trees) or significantly more mobile organisms of comparable size. Sloths may thus provide unique insights into ecological connectivity and movement ecology of wild, free-ranging animals (see Jacoby & Freeman, 2016;Schlägel et al., 2020). As discussed herein, sloths can travel ≥38 m per day but typically do so in <1 h bursts of activity and commonly within a range of <2 ha (Section II.2); they can be found at various vertical heights between the forest canopy and the ground, to which they descend once a week to defecate. The curious abundance and diversity of microbes and arthropods that take up residence within the sloth pelage raises the question as to whether the uniquely slow timescales at which sloths move, their intermittent patterns of lateral movement, together with their vertical migrations, might facilitate this phenomenon. Is there some sort of temporal resonance of ecosystem processes with sloth movement dynamics that facilitates the striking biodiversity on sloth fur? How does community diversity change as a function of the characteristic timescales of underlying assembly/dispersal processes, animal movement, and forest structure?
The movement of sloths throughout their range and up and down the canopy column may connect and disperse fur epibionts between very different ecological niches. As sloths are scattered across the tree canopy, finding, catching, and studying sloths can be experimentally challenging. However, with recent advances in Global Positioning System (GPS) tracking and remote-sensing/monitoring technology (Kays et al., 2015;Lennox et al., 2017;Neethirajan, 2017;Taylor et al., 2017;Hughey et al., 2018;Shipley et al., 2018;Williams et al., 2019;Ripperger et al., 2020), it may now be easier and more feasible to pursue continuous monitoring studies of sloths that are otherwise difficult to investigate by traditional search-and-catch methods. These capabilities may make slothsalong with their entourage of microbial and arthropod epibiontsa tractable model for exploring questions of context dependency given habitat differences across their large geographical range. Accurate time-resolved data of sloth movements in three dimensions (latitude, longitude, and altitude/elevation) are currently lacking, limiting a deeper understanding about how sloths move through the forest, their interactions with their environment and other animals, and their responses to habitat degradation or change (Pool et al., 2016;Santos et al., 2016;Brandão et al., 2019;Garcés-Restrepo, Pauli & Peery, 2019a). Data on social and habitat connectivity are critical for understanding the sources and modes of symbiont transmission. Coupling movement (spatial geo-tracking) data with real-time local environmental sensing (or time-series data) of temperature, humidity, light, etc., and with periodic biodiversity surveys of sloth fur, would provide invaluable insights into the degree of variation and environmental conditions that a sloth experiences regarding how the fur ecosystem is structured and changes.
(4) The nature and network of sloth-epibiont interactions Thus far, few studies have even attempted to simply determine the nature of the interactions between sloths and their fur epibionts. Building upon this limited knowledge, future efforts should aim to identify the symbiotic traits of each interacting organism and the selective pressures acting on those traits. The ecosystem functions of sloths within their native habitat are largely unknown, although they are believed to be an important source of long-term stable nutrients at the base of trees where they defecate (Montgomery & Sunquist, 1975). It will be important to determine through environmental sampling if algae like T. welckeri are generally limited to growth on sloths or if they can grow independently on other environmental substrates within the sloth habitat. If found environmentally, it would provide support for a model in which sloths acquire algae from the environment, and will provide a proper null model by which to assess sloth-algae coevolution as discussed in Section III.1. Other organisms are found in sloth fur, including bacteria, euglenozoans, amoebozoans, cercozoans, and alveolates (  Lincoln, 1988;Suutari et al., 2010), many of which appear not to be present in the environment around sloths (Suutari et al., 2010). The functions of these organisms in the sloth hair ecosystem are unknown but have the potential to directly impact sloth health. Sloths appear to be carriers for several arthropod-borne viruses and parasites and understanding the basis for why sloths seem not to be burdened by such pathogens may be of relevance to human health. Also unclear is the role that microbial epibionts have in facilitating host defence against pathogens in general, which has been well demonstrated in plant and pollinator systems (Liu et al., 2019).
As discussed above, the slow but unusual movement dynamics of sloths through the 'volume' of forest canopy may enable them to be hotspots of biodiversity. The stopand-go pattern of movement of sloths and their migrations of over several m/day at relatively slow speeds may make them attractive for arthropod occupancy and dispersal via hitchhiking. The slow movements of sloths may also enable them to be colonized more easily by biota like lichenous fungi (being more similar to a tree than many fast-moving animals; Supplementary Videos S1 and S2) that require low-levels of movement disturbance. Interestingly, epizoic lichens, fungi, and/or cyanobacteria have been found to grow on arthropods, specifically two species of leaf mantis in the genus Choeradodis (Lücking, Mata-Lorenzen & Dauphin, 2010) and various harvestmen arachnids (within small pits) (Machado & Vital, 2001;Proud et al., 2012;Young, Moore & Townsend Jr., 2018). Whether any of the sloth-associated arthropods carry these taxa epibiotically is unknown, but fungal-algal associations in sloth fur could potentially link sloths to arthropods and bacteria.
As primary producers, photoautotrophic algae are at the base of the food web in many ecosystems (Polis & Hurd, 1995;Segovia et al., 2015;Kohlbach et al., 2016;Brocks et al., 2017), and are likely to serve as the base of the sloth fur ecosystem as well. Algae may be dependent on nitrogen from resident fur arthropods (via faeces) or nitrogenfixing bacteria, but there are no data available on this. It is also unclear how algal growth influences the composition of the rest of the microbiome and if arthropods farm and/or consume the algae. Microbial epibionts in sloth fur may provide supporting services, including producing 'pioneer' metabolite products that provide a foundation for community development, biofilm formation, nutrient cycling, and a thriving ecosystem (McKenney et al., 2018). As a poorly studied reservoir for potentially novel microbial and genetic diversity, these hair algae/microbes may produce specialized or secondary metabolites that prevent infections, or volatiles that repel ectoparasites/predators or attract arthropods in a similar manner to how plants use volatiles to attract or repel pollinators and predators (Kessler & Baldwin, 2001;Pichersky & Gershenzon, 2002). In so doing, these natural products may play a vital role in the chemical ecology of the fur ecosystem, in shaping epibiont community structure, and in modulating sloth scent in a beneficial manner. Microbes associated with insects are known to be a source of bioactive compounds and enzymes that have biotechnological potential (Berasategui et al., 2016) and sloth microbes may ultimately be of relevance to human health and agriculture.

V. CONCLUSIONS
(1) Sloths are slow-moving, tree-dwelling, leaf-eaters in Central and South America that experience life and movement within an unusual regime of time and space, unlike any other macro-organism. With a taxonomically diverse epibiotic community residing in their fur, sloths can be considered 'mobile ecosystems' or moving islands of biodiversity. They are unique models for investigating ecological and evolutionary processes linking microbes, arthropods, animal host movement, and neotropical forest resources and structure.
(2) Sloth hair is unusual in having cracks or grooves that absorb water and support the growth of epibiotic algae, which is the basis for the distinct coloration of sloths in the wild. These primary producer algae are part of a rich multi-trophic community of microbes (including bacteria, fungi, and protists) and arthropods (including moths, beetles, sandflies, triatomine bugs, lice, mites, and ticks). Sloths are also carriers of a variety of arthropod-associated viruses ('arboviruses'). To date, how the fur ecosystem assembles, functions, and the benefits it provides the sloth host are unclear at best. (3) It is not known whether sloths have coevolved with their fur epibionts nor how these epibionts are transmitted across generations of sloths. Incomplete studies to date suggest that sloths may host unique species of algae (Trichophilus welckeri), ticks (Ambylomma geayi and Ambylomma varium), and moths (Cryptoses choloepi) that rarely if ever are found on other hosts, consistent with the idea of vertical epibiont transmission. However, this hypothesis has yet to be tested empirically, and the absence of data on environmental species pools (which would support modes of horizontal transmission) is not data of absence. Sloth-associated fungi have been found in soil and plants and over a third of fungal isolates obtained from sloth fur are identical to endophytes obtained from plants in the area, suggesting that horizontal transmission of at least some microbial epibionts may be likely. (4) Accurate animal tracking of sloths in four dimensions (latitude, longitude, altitude/elevation, and time) as they move through up, down, and across the tree canopy would provide extremely useful data for understanding the movement ecology of these mobile ecosystems as they interact with different components of the neotropical forest. Coupling movement data with real-time environmental sensing of temperature, humidity, and light (with periodic biodiversity surveys of sloth fur) would provide particularly valuable insights. (5) Experimental efforts should focus on elucidating sloth fur ecosystem functions, the nature of epibiontepibiont and sloth-epibiont interactions, and selective forces that shape this ecosystem. Understanding how fur ecosystem diversity and resilience impacts sloth health is an important practical target that could inform best practices on caring for sloths at rehabilitation facilities and zoos. Like other mammalian microbiomes, sloth fur microbiota may play an important part in immune function, and may also play a vital role in the chemical ecology of the fur ecosystem and of sloths with other individuals.

VI. ACKNOWLEDGEMENTS
We thank Michael Clear, Xia Li, Amber Horning, Margaret Campbell, and Thomas Collins for their insightful comments on early drafts. We also thank Rob Carmichael, Steve Reichling, Susan Balenger, Colin Jackson, Victoria Calatrava, and Kenny Coogan for providing feedback, Peter Marting for helpful discussions about Azteca ant and Cecropia ecology, Paul Lago and Bert Kohlmann for help identifying the scarab beetle commonly associated with sloths, Carole Moncoquet for permission to use one of her sloth videos as supplementary material, and Julio Bustamante of the Sistema Nacional de Areas de Conservaci on (Costa Rica) for issuing permit ACOPAC-INV-001-17 to S.T. Finally, we thank Judie Bronstein and other anonymous reviewers for their very constructive comments. This work was supported in part by the National Science Foundation under grants DEB-1541538 (GoLife) and DEB-1846376 (CAREER), and a sabbatical fellowship to E.F.Y.H. at the Deutsches Zentrum für Integrative Biodiversitätsforschung (iDiv) of Halle-Jena-Leipzig, Germany under grant DFG-FZT 118, 202548816 (German Research Foundation). This is the UM Center for Biodiversity and Conservation Research publication No. 012.

VIII. Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Video S1. A male Bradypus variegatus (brown-throated threefingered sloth) climbs up a tree after being released following a health check-up at The Sloth Institute in Manuel Antonio, Costa Rica. Video S2. A female B. variegatus (brown-throated threefingered sloth) climbs up a tree after being released following a health check-up (at The Sloth Institute in Manuel Antonio, Costa Rica) on her and her baby, seen clinging to her abdomen. Video by Carole Moncoquet (used with author's permission). Video S3. A female B. variegatus (brown-throated threefingered sloth) climbs up a tree after being released at night Biological Reviews 96 (2021)