Coevolution of Song and Egg Coloration: Multimodal Mating
Total Page：16
File Type：pdf, Size：1020Kb
bioRxiv preprint doi: https://doi.org/10.1101/2020.04.26.062927; this version posted April 28, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. 1 Coevolution of song and egg coloration: multimodal mating signals? 2 3 4 5 6 Silu Wang 7 University of California, Berkeley 8 Email: [email protected] bioRxiv preprint doi: https://doi.org/10.1101/2020.04.26.062927; this version posted April 28, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. 9 Abstract 10 The divergence of reproductive traits frequently underpins the evolution of reproductive 11 isolation. One of the most enduring puzzles on this subject concerns the variability in egg 12 coloration among species of tinamou (Tinamidae) —a group of birds endemic to neotropics 13 (Cabot 1992). Specifically, some tinamous lay glossy and colorful eggs while others lay less 14 colorful eggs. Here I tested the hypothesis that tinamou egg coloration is a mating signal and its 15 diversification was driven by reinforcement. For most tinamou species, the male guard the nest 16 that is sequentially visited and laid eggs in by multiple females. The colorations of the existing 17 eggs in the nest could signal mate quality and species identities to the upcoming females to the 18 nest, preventing costly hybridization, thus were selected to diverge among species. If so, egg 19 colors should coevolve with the known mating signals as the tinamou lineages diverged. The 20 tinamou songs are important mating signals and are highly divergent among species. I found that 21 the egg luminance was significantly associated with the first principle component of the song 22 variables among 31 tinamou species (after correcting for phylogenetic signal). Egg color and 23 songs could be multimodal mating signals that are divergently selected as different tinamou 24 species diverged. Mating signal evolution could be opportunistic and even exploit post-mating 25 trait as premating signals. 26 27 Introduction 28 Reproductive trait divergence is crucial for speciation because these traits frequently 29 underpin barriers to gene flow (Koski and Ashman 2016; Pfennig 2016). A very puzzling 30 reproductive trait divergence exists in tinamou egg coloration. Tinamiformes (common name: 31 tinamou) is the most species-rich order of Palaeognathae, containing 48 extant species (Figure 1) 32 (Cabot 1992; Davies 2002). Tinamous are mostly dull in plumage, but various species of 33 tinamous lay brightly colored eggs ranging from brilliant magenta, to pink, purple, turquoise, and 34 olive green (Figure 1) (Cabot 1992). Despite its evolutionary importance, tinamou egg color 35 divergence remains a mystery due to the cryptic and elusive plumage and behavior of the birds 36 (Cabot 1992; Davies 2002). 37 Two main hypotheses have been made about evolutionary mechanism of egg color 38 divergence: (1) aposematism hypothesis (Swynnerton 1916) states that eggs are brightly colored 39 to warn predators of their distastefulness; (2) mating signaling hypothesis, which predicts that 40 egg color are employed for mating recognition and attraction (Weeks 1973; Brennan 2010). The 41 aposematic function of egg coloration was thought to be unlikely, because egg predators are 42 nocturnal mammals or reptiles that prioritize chemical cues over visual cues (Skutch 1966; Cabot 43 1992; Brennan 2010). The mating signal hypothesis is plausible considering the ecology and 44 mating systems of tinamou species. In most tinamou species, males collect and incubate eggs 45 laid by multiple females (Cabot 1992; Davies 2002). An initial female is attracted by distinctive 46 songs to the nest guarded a male. If mating occurs, the female lays eggs in the nest and leaves 47 before other females come and lay more eggs that are all incubated by the same male (Cabot 48 1992; Davies 2002). The colors of the existing eggs in the nest could signal intra-specific mate 49 identity to prevent costly hybridization. Such alternative “mating channel” might be especially 50 favored when the plumage is adapted for camouflage (Cabot 1992; Davies 2002). If egg colors of 51 different tinamou species is adapted for mate recognition, reinforcement (Dobzhansky 1937; 52 Mayr 1942; Liou and Price 1994; Servedio 2000) could have driven egg color divergence among 53 different tinamou species. bioRxiv preprint doi: https://doi.org/10.1101/2020.04.26.062927; this version posted April 28, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. 54 Here I test the mating signal hypothesis of tinamou egg color divergence. If tinamou egg 55 colors are employed as a species-specific mate recognition signal, it should coevolve with other 56 mate-recognition signals as the tinamou lineages diverged. Songs of tinamous are known to be 57 important and highly divergent mating signals among species (Cabot 1992; Bertelli and Tubaro 58 2002; Laverde-R and Cadena 2014). In addition to songs, multimodality of mating signals is 59 selected for efficient mate-recognition in complex environment (Partan and Marler 1999; Rowe 60 1999; Secondi et al. 2015; Halfwerk et al. 2019). When glamorous plumage is costly due to 61 increased predation risk, alternative signal modality such as egg coloration could be favored. 62 Since the birds themselves (instead of the eggs) tend to attract nest predation (Brennan 2010), 63 egg coloration could be employed to fulfill signal multimodality, bypassing the evolutionary 64 constraints for plumage. To investigate this possibility, here I test whether egg coloration 65 coevolved with songs as tinamou species diverged. 66 67 Methods 68 To test coevolution of egg color and song in tinamous, I tested the association of songs 69 and egg colors among 32 tinamou species in R (Hothorn et al. 2008). Tinamou egg coloration 70 data was extracted from existing tinamou egg color analysis (Schläpfer 2017), which quantified 71 coloration of the eggs from 32 tinamou species (Figure 1). Since tinamou egg colors are known 72 to decay over the course museum storage, tinamou nest photos are used to best represent the 73 functional egg color (Schläpfer 2017). Briefly, for each species, the egg coloration reflected in 74 hue, chroma, and luminance in the CIELAB color space was quantified (Schläpfer 2017). 75 Song data was acquired from a previous study (Bertelli and Tubaro 2002), in which four 76 song variables from 40 tinamou species were quantified: maximum frequency (Hz), minimum 77 frequency (Hz), emphasized frequency (frequency of the note with highest amplitude in the song, 78 Hz), and bandwidth (the difference between maximum and minimum frequency, Hz). Because 79 these variables are correlated (Figure S1), I used principle component analysis (PCA) to generate 80 PC1 that captured 81.2% of the variation in song data. Four song variables were shifted to center 81 around zero and scaled to unit variance preceding PCA. 82 I tested whether egg hue, chroma, and luminance is respectively associated with song 83 PC1 among tinamous, accounting for the potential confounding phylogenetic signals. I used the 84 Tinamidae phylogenetic tree inferred with both molecular (1143 bp mitochondrial and 1145 bp 85 nuclear) and morphological characters (237 characters) (Bertelli and Porzecanski 2004), in which 86 the polytomy was resolved with additional morphological and life history traits (Bertelli 2017). I 87 first tested if there is phylogenetic signal (Pagel 1999) in each of the egg color variable with 88 phytools (Revell 2012). If no significant phylogenetic signal was identified, I ran permutation 89 test of independence between song PC1 and the egg color variable with coin package. If there is 90 significant (p < 0.05) phylogenetic signal, I first calculated independent contrasts of egg colors 91 and song PC1 with Ape (Paradis and Schliep 2019), then executed independence test. To correct 92 for multiple hypotheses testing, I conducted False Discovery Rate correction (Benjamini and 93 Hochberg 1995) to correct for three tests. 94 95 Results 96 I found a significant association between tinamou songs and egg luminance. Song PC1 97 effectively represents variation among the song variables in tinamous (Figure 2A). Significant 98 phylogenetic signal was observed in egg hue (K = 0.43, p = 0.049) and chroma (K= 0.96, p = 99 0.001), but not in luminance (K = 0.37, p = 0.17). The tinamou song PC1 was significantly bioRxiv preprint doi: https://doi.org/10.1101/2020.04.26.062927; this version posted April 28, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. 100 associated with egg luminance (Z = -2.42, p = 0.015, pcorrected = 0.045; Figure 2B), but not with 101 egg hue (Z = 0.34, p = 0.72, pcorrected = 0.75), nor with egg chroma (Z = -0.32, p = 0.75, pcorrected = 102 0.75). Apterix australis Crypturellus noctivagus Crypturellus