On the evolution of mimicry in avian nestlings

Abstract Batesian mimicry (BM), where a nontoxic species resembles a toxic species with aposematic coloring, has been recently described for a Neotropical species of the suboscine passerine (Laniocera hypopyrra). Understanding the order and series in which these characteristics evolved is unknown and requires character information from closely related taxa. Here, we trace the origin of mimetic traits and how they evolved by examining antipredator characteristics using images and other field‐collected trait data from nest and nestlings along with data available in the literature for the Laniisominae clade and closely related taxa. We found that morphological modifications of the downy feathers appeared first in the broader clade leading to the Laniisominae clade followed by further morphological and behavioral characteristics within the Laniisominae clade leading to the full BM. Images of nestlings in the Laniisominae and closely related clades demonstrated the extent of antipredator and camouflage characteristics. We found a complex set of behavioral and morphological traits in this clade for reducing predation from hiding to camouflage to mimicry. We further propose the evolution of two distinctive mimicry strategies in the Laniisominae clade: (1) Batesian Mimicry, as described above and (2) Masquerade, resemblance to inedible objects commonly found in their local environment. This complex set of antipredator traits shed light on the diversity of antipredator characteristics in avian nestlings, particularly in neotropical areas where the avian diversity is highest. Unfortunately, there are hundreds of species in the neotropics that lack basic natural history information on nesting traits, and therefore, we are likely missing critical information on the diversity of antipredator characteristics across avian nestlings.


| INTRODUC TI ON
Understanding how animals' morphology and behavior change in response to selection from predation remains an important question for evolutionary theory (Ruxton et al., 2018). For birds, nest predation is an important factor affecting avian fitness (Martin, 1993;Ricklefs, 1969;Robinson et al., 2000). Nest predation is not homogeneous for all nestlings and may vary depending on the habitat or on avian characteristics. For example, predation is thought to be higher at lower latitudes and for birds that build open nests (e.g., cup-shaped; Griebeler et al., 2010;Martin, 1995). In areas with high predation rates, birds exhibit multiple proximal factors that include behavioral, morphological, and physiological traits thought to reduce predation. For example, birds can change their nest location or nest type, as closed nests (e.g., globular or cavity; Griebeler et al., 2010;Martin, 1995) or those located lower in the canopy are thought to be harder to find by predators (Hoset & Husby, 2019;Major et al., 1994;Storaas, 1988). A reduction in the nesting period or number of foraging trips away from the nest limits the exposure time and probability of a predator finding the nest through observing adults flying to the nest (Ghalambor & Martin, 2002;Martin et al., 2007;Robinson et al., 2010;Zanette et al., 2011). Compensatory traits such as increased numbers of nesting attempts or longer nesting seasons may also have evolved in response to predation pressure (Roper et al., 2010).
A highly specialized evolutionary strategy to avoid predation is Batesian mimicry, a phenotypic and sometimes behavioral trait where a nontoxic species resembles a toxic species with aposematic coloring (i.e., true warning coloration; Mallet & Joron, 1999).
The origin of Batesian mimicry (BM) has long been controversial in evolutionary biology. The debate has included two distinct hypotheses: It has been envisioned either as a cumulative process of trait changes that result in BM (Leimar et al., 2012;Mappes & Alatalo, 1997) or as a multistep process that involves major mutations resulting in BM. The latter could be also seen under the modular evolution hypothesis, where the traits are treated as units that change step by step in a modular fashion (Eble, 2005). This model has been used to explain the morphological and behavioral diversity present in other avian groups (e.g., the enigmatic feather patterns and behaviors of birds of paradise; Scholes, 2008).
One hypothesized avian case of Batesian mimicry (BM) in nestlings is that of Laniocera hypopyrra, a Neotropical bird in the family Tityridae found in subtropical and tropical lowland forests across South America. The nestlings of L. hypopyrra possess bright orange elongated barbs that mimic caterpillars from the Megalopygidae family (D'Horta et al., 2012; Figure 1), which inhabit the same habitats. These caterpillars are well known for their toxicity (Deml & Epstein, 2001;Dyar & Morton, 1895;Hossler, 2009;Lamdin, 1976;Sánchez et al., 2019), and can be lethal to humans (Pinson & Morgan, 1991). The hypothesis that nestlings of L. hypopyrra are Batesian mimics is supported by morphological and behavioral traits that enhance the resemblance of nestlings to toxic caterpillars (Londoño et al., 2015). Interestingly, this specific combination of traits in L. hypopyrra nestlings are not observed in other closely related species (D'Horta et al., 2012;Londoño et al., 2015); in particular, the nestlings do not move or beg in the nest, typical of other nestlings. Furthermore, they move their heads from side to side in a motion that resembles the movements of a caterpillar when they perceive a potential predator (e.g., humans approaching the nest). The morphological traits that enhance their similarity to the Megalopygid caterpillars include increased density, coloration change, and structural modification of the downy nestling feathers (Londoño et al., 2015). The combination of morphological and behavioral traits likely evolved through a complex trajectory of antipredator characteristics resulting in Batesian mimicry (Londoño et al., 2015). Our goal is to understand in what order these mimetic traits evolved by reconstructing the ancestral character states of L. hypopyrra and closely related species to shed light on the number of antipredator characteristics found across this group, and to illuminate the way that this mimicry evolved. For example, do we see evidence of the more commonly known antipredator traits for birds (e.g., a decrease in the number of trips to the nest), or did an entirely different evolutionary trajectory occur in these species? How widespread are mimetic traits in this group and in what order do antipredator traits evolve?

F I G U R E 1
Here, we compile morphological and behavioral characteristics of the nestlings from L. hypopyrra and close relatives to conduct an ancestral character state reconstruction. While some of these morphological features have been known for some time, the nestling behaviors have not been described for many of these species.
We combine data from our fieldwork, and from unpublished and published information on morphological and behavioral traits to track evolutionary trajectory of the antipredator traits observed in this clade.

| ME THODS
We selected 11 representative species from nine closely related genera based on a recent suboscine passerines phylogenetic reconstruction of this larger clade (Harvey et al., 2020). Species were selected based on the presence of nestling traits in the literature or from our field work. Nestling characteristics are rarely identified and cataloged due to the difficulty in finding nests and collecting these data. In this dataset, we have a single represtative for each genus and this dataset represents the largest set of nestling characterisitics across these clades. From the Laniisominae clade, we included a representative from Laniocera, Laniisoma, and Schiffornis, from its sister clade, Tityrinae; we included the genera Tityra, Pachyramphus and Iodopleura, and from the tribe Onychorhynchini, we included Onychorhynchus, Terenotriccus, and Myiobius, which is the sister clade of the genus Oxyruncus (Harvey et al., 2020;Tello et al., 2009; Figure   S1). This sampling represents all but a single genus within this larger clade. We used the species Piprites chloris a sister species of the family Tityrinae as the outgroup. Most of these species live in sympatry and share the same community of nest predators.

| Nesting traits
A total of twelve traits were compiled from published and unpublished data for 11 bird taxa ( Table 1; Table S1) representing morphological (modified elongated, dendritic, dense, and juvenile down feathers), behavioral (reduced begging, no movement, and behavioral mimicry), and compensatory traits (length of the nestling period, feeding rate, nest type, nest location, and the number of trips to the nest; Table 1). We also collected pictures of nestlings to examine nestling camouflage found across this clade. We collected missing data from field observations for Schiffornis stenorhyncha, Pachyramphus cinnamomeous, Tityra semifasciata, Onychorhynchus coronatus, Terenotriccus erythrurus, and Myiobius atricaudus. These data were collected, using standardized methodologies found across studies, during a long-term avian nesting fieldwork project conducted by GAL in Peru and Colombia (see details in the supplementary methods).

| Ancestral character state reconstruction
To reconstruct the ancestral character states, we obtained a phylogenetic tree from Harvey et al. (2020), which included all the species in this study, and pruned it to our 11 taxa (see Figures S1-S4).
This topology does not differ from other studies focused on species within these groups (Prum & Lanyon, 1989;Tello et al., 2009).
We further downloaded 1000 trees from birdtree.org (Jetz et al., 2012) to examine uncertainty across the distribution of trees. The major clades were identical across the 1000 trees apart from the placement of Oxyruncus. These taxa were placed in different clades but never within the Laniisominae or Tityrinae clade. Furthermore, running the analysis on all 1,000 trees did not change the results of the major nodes represented in the results (data not shown). We note that the species represented in our tree as Schiffornis stenorhyncha has been split into five species (Chesser et al., 2013;Nyári, 2007;Remsen et al., 2022). However, these species are not included as independent branches in the tree, rather they are presented as a single branch (Schiffornis stenorhyncha), the overall tree topology with relationship to the other genera is not changed and the nesting traits known for the five species do not vary.
To determine the trend and order of the evolution of these traits, we first determined the best fit model of evolution for each character and, secondly, ran an ancestral character state reconstruction to determine the character state at each ancestral node. With the character state identified at each node, we then determined the progression of character states leading to Batesian Mimicry in this group. We used Geiger (Pennell et al., 2014) and Phytools (Revell, 2012) libraries in R statistical packages to model the evolution of characters and reconstruct the ancestral character states. The phylogenetic tree was visualized and generated using ggtree (Yu et al., 2017) and plotrix (Lemon, 2006) packages in R statistical software (R Core Team, 2021).

| Discrete character states
The nine discrete data included both binary (1,0; present, absent) and multistate (1,2,3) characters (Table 1). We evaluated three models to determine which was the best fit for each character. The

| Continuous character states
We first determined which model of evolution had the best fit for the three continuous characters using four different models: (i) Early Burst ( ues that maximize the probability of the data on the tree. We examined the estimated ancestral states to determine whether they were higher or lower than the current states to understand how these characters have changed overtime.

| RE SULTS
We were able to collect data for most traits for all of the taxa either from the literature or from the field ( Table 1). Iodopleura isabellae presented the largest challenge with the fewest characters in the literature. This species nests in the canopy, which makes finding and monitoring nests particularly difficult and likely explains the lack of nesting data.

| Down feather modification
Schiffornis turdina (Skutch & Eckelberry, 1969) and S. stenorhyncha nestlings were completely covered with a dense brownish down, with a few elongated dendritic barbs of the same color (Figure 2), sharing with L. hypopyrra the density and elongation of the down barbs. For T. semifasciata, we obtained information from Skutch and Eckelberry's (1969) observations that describe the presence of dendritic down for nestlings with "fairly long light gray down" on head, wings, and back that did not entirely cover the pink skin. By contrast, Pachyramphus polychopterus (Skutch & Eckelberry, 1969) and

| Nestling behaviour
We know that the behavioral mimicry in L. hypopyrra is exhibited only during a portion of the nestling period (Londoño et al., 2015).
During most visits to the nest, S. stenorhyncha nestlings stood still exhibiting no movement and did not beg for food, which was accompanied by lowering the head toward the belly when the researchers approached the nest. Similarly, Skutch and Eckelberry (1969) describe that when they approached and shook the nest of S. turdina or held a hand above the nestlings, they did not lift or gape their mouth as many young nestlings do, they remain motionless. These behaviors were not observed during the whole nestling period of S. stenorhyncha. Therefore, it is possible that some of these behavioral characteristics are only present during a portion of the nestling period similar to L. hypopyrra (Londoño et al., 2015). Skutch and Eckelberry (1969)

| Traits within other members of the Laniisominae clade
For all of the nine discrete traits, the Equal Rates (ER) model was found to be the best fit, and for the three continuous traits, the Brownian Motion (BM) model was found to be the best fit for each character. Through ACR of traits, we were able to assess that the  (Caro, 2014;Skelhorn et al., 2010), where the mimetic organism resembles objects of no particular interest to the predator, such as rotten fruit (Snow, 1982). In this case, the nestlings of Schiffornis stenorhyncha resembles a fungus found in the area (Figure 1). This suggests that mimicry may have evolved in the ancestor of the Laniisominae The risk of predation is higher in cup nesters than for those who nest in cavities or globular nests observed in the sister clades (Arendt, 1997;Martin & Li, 1992;Oniki, 1979;Ricklefs, 1968;Wilcove, 1985). L. hypopyrra and S. stenorhyncha are cup nesters, which likely increased predation pressure on those nests and may be the reason further antipredator traits were selected for in Laniocera and Shiffronis including the reduction in the number of visits to the nests and feeding rate (Martin et al., 2000). The slow feeding rate could also be affecting the growth rate (Crossner, 1977;Ricklefs, 1968), increasing the nestling period (Richner et al., 1989). The main constraint to tracking the evolution of these traits in Laniocera is the lack of natural history information from closely related species. In our sample, we have a single representative across these genera making this a relatively small sample size. This limitation is likely the reason we found that Brownian Motion was the best fit model for the contiuous data. Brownian Motion suggests neutral evolution; however, we argue that these extreme rare traits (e.g., bright orange coloring and non-begging nestlings) are more likely the result of selection and our result is due to the small sample size. For example, we know that the sister taxon to L. hypopyrra, Laniisoma elegans, has some of the morphological traits exhibited by L. hypopyrra, including the bright orange coloration and structural downy feather modifications, but whether nestlings display behavioral mimicry is unknown (see details on Londoño et al., 2015). Some information on the downy characteristics and coloration is known for the sister genus Schiffornis (Sandoval-H et al., 2017;Skutch & Eckelberry, 1969). For example, the downy feather of Schiffornis veraepacis and S. stenorhyncha are brownish-gray and more abundant than that of nestlings of the majority of passerine birds (Sandoval-H et al., 2017;Skutch & Eckelberry, 1969), and the latter has dense down composed of elongated dendritic barbs, suggesting that pe-  Willis and Oniki (1988), the old nestling of I. pipra has the color and the irregularity of the dorsal plumage, resemble a scaly mass of bark or a bulky mass of debris on the flattened nest, so that it was not obvious when it sat quietly, as it did much of the day. This behavior and coloration of the old nestling of I. pipra match the coloration and feather pattern of old nestlings from the genus Nyctibius that uses cryptic traits to reduce predation. Additionally, Idopleura spp. has a very small nest for their size and presumably one egg clutch size (Ingels & Vinot, 2010;Snow, 1982;Whittaker & Kirwan, 2008;Willis & Oniki, 1988), traits associated with high predation.
Furthermore, the behavioral trait evolved in a similar way where Laniisominae sister clade nestling exhibit a hiding behavior inside the nest, while the nestlings of S. stenorhyncha and S. turdina (Skutch & Eckelberry, 1969) exhibit a no movement behavior and the L. hypopyrra nestlings exhibit a complex head movement when in danger.
Taken together, these results suggest a complex set of behaviors and traits in this clade limiting predation from hiding and camouflage to Masquerade and Batesian mimicry, expanding our ideas of the possibilities and lability of antipredator behaviors in avian nestlings, mainly in neotropical areas where the avian diversity is the highest.
It is likely that the evolution of mimetic and antipredatory strategies observed in this group occurred under the selection pressure of a community of predatory species, rather than a single predator species that impose a high predation pressure (Schmidt, 1958). Yet, this may only be important for species that build open nests (e.g., cup and platforms), as few nest predators have access to close nests (e.g., cavity and globular).
Unfortunately, there are hundreds of bird species in the neotropics that lack basic natural history information on nesting traits (Xiao et al., 2017). Hence, we encourage naturalists to explore the neotropical forest in search of the unknown nests that could uncover amazing traits such as the adaptive coloration and diverse antipredatory strategies observed in this group.

ACK N OWLED G M ENTS
More than 220 field and laboratory assistants were essential in obtaining the data used in this paper. In particular for this study, we Research Award and the Katherine Ordway Foundation.

CO N FLI C T O F I NTE R E S T
We do not have any conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All the raw data are avalible on the Appendix S1 tables.