INTRODUCTION

Bird vocal behavior gives rise to a highly effective acoustic communication system (Marler and Slabbekoorn 2004). Song, the most complex acoustic signal, is used for mate attraction and selection, and in male–male competition, including territory defense (e.g., Searcy and Andersson 1986, Nowicki and Searcy 2004). Given these important roles of song, it is likely that vocal quality is subject to strong selective pressures (e.g., Marler and Slabbekoorn 2004). Singing birds may express quality by a variety of acoustic features of the vocal repertoire, among which are rapid temporal aspects such as trill and modulation rates (e.g., Podos 1997, Draganoiu et al. 2002, Schmidt et al. 2008, Suthers et al. 2012). Production of high trill rates encompasses rapid switching between sound production and silent periods, and often includes rapid frequency modulation rates (e.g., Suthers and Goller 1997, Suthers and Zollinger 2008, Riede and Goller 2010, Elemans 2014). If these features are linked to neuromuscular control, the muscle kinetics of the involved motor systems must be sufficiently fast.

The oscine syrinx is a bipartite structure, consisting of 2 distinct sound generators, one located in each bronchus. Each sound source is independently controlled by at least 4 pairs of intrinsic muscles, ventral and dorsal tracheobronchial and ventral and dorsal syringeal muscles (Figure 1; e.g., Larsen and Goller 2002, Düring et al. 2013). Song features arise as the result of the interaction between passive dynamic properties of the sound generating labia with physical parameters (airflow and driving pressure) as well as direct muscular control (e.g., Suthers and Zollinger 2008, Riede and Goller 2010, Goller and Riede 2013, Elemans 2014). Activity of the syringeal muscles permits direct control of airflow and labial tension and therefore affects acoustic parameters such as frequency and amplitude and facilitates rapid modulation of both (e.g., Vicario 1991; Goller and Suthers 1995, 1996; Larsen and Goller 2002; Riede and Goller 2010; Goller and Riede 2013).

FIGURE 1.

Schematic view of European Starling syrinx in ventral view (A) and dorsal view (B). Abbreviations: dS = dorsal M. syringealis; dTB = dorsal M. tracheobronchialis; dS = dorsal M. syringealis; ST = M. sternotrachealis; vS = ventral M. syringealis. M. sternotrachealis not illustrated on the dorsal view (modified from Uchida et al. 2010). In (C) a cross-section through the syrinx of a European Starling shows the immunohistochemical reaction using anti-fast antibody MY32. Red fibers indicate a positive reaction for fast fibers. White (unreacted) fibers are presumptive superfast fibers. Ventral is to the bottom of the page. Scale bar = 500 μm

Sexual dimorphism in singing behavior is accompanied by differences in both the central nervous system (e.g., Nottebohm and Arnold 1976, Cooke et al. 1998, Wade and Arnold 2004) and in the syrinx (e.g., Wade and Buhlmann 2000, Wade et al. 2002, Riede and Goller 2010). Several areas of the central song control circuitry are larger in males than in females in the Island Canary (Serinus canaria) and the Zebra Finch (Taeniopygia guttata) (e.g., Nottebohm and Arnold 1976). In the Zebra Finch, in which only males sing, the syrinx of males weighs approximately twice as much as that of females (despite minimal body size dimorphism); this difference is attributed in large part to the increased muscle mass in males (Wade and Buhlman 2000). Similarly, the syrinx of male European Starlings (Sturnus vulgaris) is larger than that of females, and the greater muscle mass is thought to be correlated with the ability to produce more complex song. Female European Starlings demonstrate smaller vocal repertoires and sing at slower rates than males (Prince et al. 2011).

Superfast muscle fibers are known from sound-producing organs in species as diverse as fishes (e.g., Kéver et al. 2014), snakes (e.g., Rome et al. 1996), birds (Elemans et al. 2008, Fuxjager et al. 2016), and mammals (e.g., Elemans et al. 2011). These fibers contract faster than typical vertebrate “fast” muscle and display a trade-off between force generation and speed (Rome 2006). In songbirds, superfast fibers have been shown to produce positive work up to 250 Hz in European Starlings and up to 200 Hz in male Zebra Finches (Elemans et al. 2008), placing them among the fastest vertebrate skeletal muscles (e.g., Rome et al. 1996, Rome et al. 1999). Uchida et al. (2010) described both superfast and fast fibers in the European Starling syrinx. Interestingly, whereas there was no sexual dimorphism in fiber type composition (Uchida et al. 2010) and in contraction rate in the European Starling, the muscles of female Zebra Finches were significantly slower than those of males (Elemans et al. 2008). While female European Starlings sing occasionally, female Zebra Finches do not. This raises the question to what degree fiber type composition and resulting contraction speed are related to singing ability and activity in different oscine species. We address this question here by sampling fiber type composition of syringeal muscles across several oscine families. We find that sexual dimorphism is uncommon and explore to what degree body size relates to muscle fiber morphology. Some of these results have been previously presented in abstracts (Uchida et al. 2009, Christensen et al. 2014).

METHODS

Tissue Preparation and Histology

Immediately following an overdose with isoflurane, syringeal tissue was obtained by removing the portion of the trachea and bronchi with the syringeal muscles intact. The syrinx was then mounted with the cranial aspect facing up on a 2 × 2 cm piece of cork using 5% gum tragacanth. Blocks were flash-frozen in isopentane, cooled in liquid nitrogen to approximately –150°C and subsequently stored in an ultracold freezer at –80°C. Serial cross sections were obtained by taking 10–12 μm sections using a cryostat (Tissue-Tek II, Microtome/Cryostat, models 4551 and 4553, respectively; Ames Division, Miles Laboratories, Elkhart, Indiana, USA) set at –20°C and mounted onto microscope slides. Sections were collected along the entire length of the syrinx to ensure that the largest cross-sectional area could be identified.

Muscle fiber types were identified using a series of histochemical techniques described in McFarland and Meyers (2008) and Uchida et al. (2010). Serial sections were reacted for myosin ATPase using acidic (pH 4.2–4.3) pre-incubations. In addition, muscle sections were reacted with anti-slow antibody ALD58 (Hybridoma Bank, University of Iowa) and anti-fast antibodies MY32 (Sigma Chemical Company, St. Louis, Missouri, USA) and EB165 (Hybridoma Bank, University of Iowa) as described in Meyers and Stakebake (2005). Primary antibody slides were incubated in a humid chamber for 2 hr at 25°C, rinsed with phosphate buffered saline, incubated with secondary antibody, and stained with a peroxidase substrate system (IMMpact AEC Red kit, Vector Labs, Burlingame, California, USA). Acidic myosin ATPase and anti-slow antibody ALD58 react positively for slow muscle fibers. Alkaline myosin ATPase and anti-fast antibodies MY32 and EB165 give positive reactions for fast muscle fibers. Superfast fibers were identified by negative reactions with both ALD58/MY32/EB165. Acidic and alkaline ATPase often showed an intermediate or moderate reaction staining (Figure 2).

FIGURE 2.

Serial sections showing muscle fiber staining profiles from a ventral tracheobronchealis (vTB) of a male European Starling. Fibers denoted by arrows are smaller, fast fibers whereas the fiber indicated by an asterisk represents the larger, superfast muscle fibers. Acid preincubation (pH 4.3) reacts darkly for slow-twitch fibers; no positive reaction was found in any syringeal muscle. Alkaline (pH 10.4) preincubation reacts positively for fast-twitch muscle fibers as shown by the arrowed fibers. The antibodies ALD58 and MY32 react positively with slow and fast muscle fibers, respectively. Note the uniformly negative ALD58 reaction indicating no slow fibers, and the positive reaction with the smaller fibers in MY32. Scale bar = 100 μm.

Reacted whole syrinx cross-sections and appropriate scale bars were photographed using a digital camera (Rebel T1i EOS; Canon, Tokyo, Japan) mounted on a microscope (Axioskop 40/40 FL; Carl Zeiss, Oberkochen, Germany) at 10 or 20×. Graphics-editing software (Photoshop CS2; Adobe Systems, San Jose, California, USA) was used to merge overlapping photographs into a complete photomontage of the entire muscle. The photomontage was printed and taped together to allow for analysis and quantification.

RESULTS

The syringeal muscles of all species were composed of 2 distinct fiber types, fast and superfast (Figures 1, 2, and 3). Fast fibers reacted like type IIA muscle fibers, with positive reactions to anti-fast antibody MY32 and negative reactions with anti-slow antibody ALD58 and acidic pH pre-incubations. The superfast fibers did not react like any typical type I, IIA, IIB, or tonic muscle fibers. These fibers had a negative reaction with both MY32 and ALD58, and a non-positive “intermediate” or “gray” reaction with acidic pH pre-incubations (Figure 2). These atypical reactions were previously described by Uchida et al. (2010) for the avian syrinx and DelGaudio et al. (1995) for the mammalian larynx.

FIGURE 3.

Representative syrinx cross-sections from males and females of 3 of the species examined in this study. Sections are reacted with MY32, which reacts the fast fibers red, and leaves superfast fibers unreacted. ZF = Zebra Finch, ESt = European Starling, BhC = Brown-headed Cowbird. All images are shown to scale. Scale bar = 500μm.

Syrinx Cross-Sectional Areas

Cross-sectional area measurements revealed that in all species, males had a larger syrinx than females (Figure 3). A 2-way ANOVA with sex and species as factors revealed highly significant effects of species and sex and no significant interaction (species: F = 16.008, p < 0.001; sex: F = 27.318, p < 0.001; species × sex: F = 1.48, ns; Table 1). Cross-sectional area increased with increasing body size. Body mass differences explain 91% and 88% of the observed variation in area in males and females, respectively (Figure 4).

TABLE 1.

Cross-sectional area (CSA) (mm²) for each species.

FIGURE 4.

Plot of the relationship between the average syrinx CSA (in mm²) vs. body mass for the males and females in this study. Muscle CSA increased with increasing body size, with males typically larger in both measurements. Linear regression: Males: diameter = 1.71 + 0.134 × mass, F = 85.6, p < 0.0001, r² = 0.91. Females: diameter = 0.651 + 0.1 × mass, F = 57.4, p < 0.0001, r² = 0.88.

The number of syringeal muscle fibers scaled positively with body size in males (superfast: F = 26.7, p = 0.0009, r² = 0.77; fast: F = 17.3, p = 0.003, r² = 0.68) and females (superfast: F = 40.6, p = 0.0002, r² = 0.83; fast: F = 7.98, p = 0.022, r² = 0.5). Interestingly, the superfast fiber counts appear to be more tightly linked to body mass than fast fiber counts. This is particularly pronounced in females.

Syrinx Muscle Fiber Diameters

In all species, both males and females had superfast fibers that were larger in diameter than the fast fibers. A 3-way ANOVA with sex, species, and fiber type as factors indicated significant effects of all factors (species: F = 6.075, p < 0.001; sex: F = 12.102, p < 0.001; fiber type: F = 154.6, p < 0.001; all interactions are not significant). Differences between fiber types are fairly pronounced in all species (Table 2). For example, the mean diameter of superfast fibers in male and female Brown-headed Cowbirds was about 32 μm and 23 μm, respectively, and mean diameter of fast fibers was 17 μm and 16 μm in males and females. In most species, post-hoc LSD tests revealed that superfast and fast fibers were not significantly different between males and females (Table 2). The most notable exceptions were found in Zebra Finches and Bengalese Finches, where mean diameter of superfast fibers was significantly larger in males (Zebra Finch: p < 0.001; Bengalese Finch: p < 0.05). In addition, mean diameter of fast fibers was larger in male Bengalese Finch (p < 0.05) and Yellow-headed Blackbird (p < 0.01).

TABLE 2.

Diameter (μm) of fast and superfast fibers for each species.

The diameters of the 2 fiber types showed a different relationship with body size. No statistically significant relationship was found in fast fibers (Figure 5A; r² = 0.19, p = 0.16 for males and r² = 0.36, p = 0.064 for females), whereas the superfast fiber diameter in males and females increased significantly with body mass (Figure 5B; r² = 0.47, p = 0.028 for males, r² = 0.45, p = 0.033 for females). However, there is substantial variation in fiber diameters beyond the relationship with body mass. For example, mean superfast fiber diameter was 27 μm in male Red-winged Blackbird, whereas that from male Zebra Finch was 26 μm.

FIGURE 5.

Plot of the relationship between the mean diameters of (A) fast fibers vs. body mass, and (B) superfast fibers vs. body mass. Males typically showed larger diameter fibers than females.

Fiber Type Composition of Syringeal Muscles

Our initial 2-way ANOVA using data from all species with species and sex as factors yielded a significant effect of both on superfast fiber percentage and a significant interaction (species: F = 9.595, p < 0.001; sex: F = 34.616, p < 0.001). We further explored whether the 2 species with distinct sex differences in superfast fiber type percentage (Bengalese Finch and Zebra Finch) were driving the significant effect of sex as a factor. For the data set of 8 species (Bengalese Finch and Zebra Finch removed), sex no longer is a significant factor (sex: F = 0.724, n.s.), indicating that a clear sex difference in fiber type composition only occurred in Bengalese Finch and Zebra Finch (Figure 3, Table 3). Post-hoc LSD tests showed male superfast percentages significantly greater only in Bengalese Finch (p < 0.001) and Zebra Finch (p < 0.01). When corrected to reflect the larger size of the superfast fibers, a comparison of male vs. female superfast percentage was again most apparently different for Bengalese Finch and Zebra Finch (Figure 6). In these 2 species, superfast content was distinctly lower in females than in all the other species. Additionally, Zebra Finch male superfast content was the highest of all species.

TABLE 3.

Total fiber numbers, superfast fiber percentages, and mean superfast fiber percentages for each species.

FIGURE 6.

Plot of the relationship between the diameter-corrected male superfast fiber percentage vs. female superfast fiber percentage. Most species are clustered near the line of equal superfast contribution in both sexes, with the exception of Zebra Finches and Bengalese Finches displaying a markedly lower percentage in females compared to that in males. Zebra Finch males also have the highest percentage of all species, whereas male Bengalese Finches fall within the distribution of males in the other species. AfS = African Silverbill, BF = Bengalese Finch, BhC = Brown-headed Cowbird, BSp = Brewer's Sparrow, ESt = European Starling, HSp = House Sparrow, RwB = Red-winged Blackbird, WSp = White-crowned Sparrow, YhB = Yellow-headed Blackbird, ZF = Zebra Finch.

In order to investigate whether superfast fiber content drives overall cross-sectional area, we regressed the residuals of the relationship between body mass and cross-sectional area (from Figure 4) on the diameter-corrected superfast percentage values (Figure 7). The weak positive relationship in males is not statistically significant (F = 3.4, p = 0.1, r² = 0.30), no relationship was found in females (F = 0.08, p = 0.78, r² = 0.01), and excluding Zebra Finch and Bengalese Finch also did not result in a significant linear relationship (F = 1.94, p = 0.21, r² = 0.24). The residual values for Brown-headed Cowbird and Zebra Finch are notably high. Unlike in the case of Zebra Finch, this large size cannot be attributed to very high superfast fiber content in Brown-headed Cowbird and must therefore reflect a greater contribution by fast oxidative fibers. The value for male House Sparrow is the lowest residual despite a medium superfast percentage, indicating a smaller contribution by fast oxidative fibers than in the other species.

FIGURE 7.

Residual syrinx CSA × mass (from Figure 4) vs. the diameter-corrected superfast fiber percentages shows a positive relationship in males (A) but not in females (B). Species abbreviations as in Figure 6.

DISCUSSION

In this study we investigate the composition of syringeal musculature in males and females of 10 songbird species. These species differ greatly in the extent of their sexual dimorphism in singing behavior, ranging from highly dimorphic singers in which males sing and females do not (e.g., Zebra Finch, Bengalese Finch, Brown-headed Cowbird, Brewer's Sparrow, African Silverbill) to those species where females at least occasionally sing (e.g., White-crowned Sparrow, European Starling, House Sparrow) or have different song-like vocalizations (e.g., Yellow-headed Blackbird, Red-winged Blackbird). The results show that the syringeal muscles of both sexes are heterogeneously composed of superfast and fast muscle fibers in all species examined, as previously described for European Starling (Uchida et al. 2010). Although there is variation between species in detailed muscle composition, the most striking result is low superfast fiber content in Zebra Finch and Bengalese Finch females.

Differences in cross-sectional area of the syrinx between species were largely explained by body mass. A clear sex difference independent of mass differences is indicated by the 3-fold higher intercept of the linear regression in males. A larger muscle mass in males has been measured consistently, and is at least in part regulated by androgens (e.g., Häcker 1900, Luine et al. 1980, Springer and Wade 1997, Wade et al. 1999). Variable levels of testosterone at the time of measurement may also explain some of the substantial inter-individual variation. The larger muscle mass in males may also reflect increased use of the syringeal muscles for vocal behavior. In many species in which females also sing regularly, males still sing much more frequently and loudly. The specific mechanisms by which higher muscle mass supports high vocal output, while often retaining high stereotypy of song, are not known.

We document substantial variation in fiber diameters of both fiber types between the different species and large variation within some species. In the syrinx, superfast fibers within a sex were found to be larger in diameter (max. diameter = 36 μm) when compared to fast fibers (max. diameter = 21 μm). Thus, superfast fibers seem to be within the normal range of avian fast fibers (see Kaplan and Goslow 1989), whereas the fast fibers of syringeal muscles are smaller than those found in avian leg muscles (see Suzuki and Tamate 1979, McFarland and Meyers 2008), though comparative data are minimal. Interestingly, the superfast fibers, which contract at rates reaching 250 Hz, are, in some cases, double the size of fast fibers. Perhaps this increased diameter compensates for otherwise low force production, which is likely to result from the trade-off between force and speed (Rome et al. 1996, Rome and Linstedt 1998, Rome et al. 1999, Rome 2006). Individual muscle size is limited by constraints of overall body size, but it is unclear whether the size of individual structures, such as the syrinx, also influences the maximal diameter of muscle fibers. The syrinx is situated in the interclavicular air sac and needs to move freely without contacting the sternum ventrally. Such restrictions may have exerted selective pressure for fairly small fiber diameters, while maintaining fine control through sufficient numbers of fibers and thus motor units. Although syringeal muscle fiber diameters do scale positively with body mass (Figure 5), there is little comparative data to suggest that this is a trend across muscles in a variety of species. Furthermore, it is well known that testosterone can increase muscle fiber sizes up to 50% (e.g., Volek et al. 1999, Sinha-Hikim et al. 2002) as can resistance training with protein supplementation (Andersen et al. 2005). Although male superfast fibers tended to be larger than those of females in all cases, this difference was only significant in Zebra Finch and Bengalese Finch. Similarly, fast fiber diameters were only significantly different in Bengalese Finch and Yellow-headed Blackbird, but the trend was found in all species. Greater sample sizes would likely establish a distinct sex difference. How the differences in fiber size affect vocal behavior is unclear, because the biomechanics of vocal control are insufficiently understood (e.g., Goller 2016).

The fiber type composition of syringeal muscles is marked by a high percentage (>50%) of superfast fibers. Whereas this high proportion of superfast fibers is found in males and females of most species, in Zebra Finch and Bengalese Finch the composition is sexually dimorphic. Although it is tempting to attribute this to the lack of song in females of these species, this behavioral difference cannot explain the absence of strong sexual dimorphism in other species in which females do not sing. In several other species of the set investigated here, females are not known to sing, yet their fiber type composition is not strongly dimorphic. So, why do we find such a distinct sexual dimorphism in fiber type composition only in these 2 species?

Superfast muscle content may be important for controlling the syringeal valves during nonvocal behaviors. In this case the low percentage of superfast fibers in female Zebra Finch and Bengalese Finch is rather puzzling and an explanation for why there should be less need for rapid respiratory control in the females of these 2 species than in all the other species is not obvious. Both Zebra Finch and Bengalese Finch are domesticated strains, and a possible relaxation of selective pressures on rapid valve control might have resulted from the domestication process. Investigations of the syringeal muscles of wild Zebra Finches and the White-backed Munia (Lonchura striata), the presumed parent species of the Bengalese Finch, would provide clarification of this possibility.

Alternatively and more likely, the sexual dimorphism may be related to the entire vocal repertoire of each sex, rather than just singing behavior or the lack thereof. If a relaxation in selective pressures to achieve high contractile speed of the syringeal muscles occurred, then we can expect that rapid control by syringeal muscles is not required to produce the call repertoire. Consistent with this interpretation is the fact that in female Zebra Finch syringeal muscles are not very active during distance call and soft call production (Goller, personal observation) and that these vocalizations do not contain acoustic features, such as fast frequency and amplitude modulation, whose production would require rapid muscle control. However, the situation is less clear in female Bengalese Finch, which generate trilled calls that may require muscular control. Detailed investigation of the role of syringeal muscles in the production of the vocal repertoire of Bengalese Finch females could clarify this issue.

The fact that there is no marked sexual dimorphism in fiber type composition in the other species most likely indicates that vocal repertoires are composed of calls and song that require contributions of rapid muscle contractions. Song production by both sexes is thought to be ancestral (Odom et al. 2014), and female song is more common in tropical species (Slater and Mann 2004). Most females possess complex call repertoires; the female Streak-backed Oriole (Icterus pustulatus) sings more than the male (Price et al. 2008). In addition, female traits (e.g., singing) can evolve rapidly and independently (Webb et al. 2016) and be under weakened selective pressure. One idea may be that weak selection may cause more variable content of superfast fibers in females across different species. However, if the development and maintenance of superfast fibers in muscles incurs a greater cost, their presence in females may also be under some positive selective pressure. Evidence for a potentially substantial cost of maintaining syringeal muscles may be seen in the fact that overall syrinx weight is increased by circulating testosterone (Luine et al. 1980), syringeal fiber size is increased by estrogen-blockers (Martin and Veney 2008), and superfast fiber percentages are increased by testosterone (Allred et al. 2011).

It is therefore likely that vocal behavior of males and females requires rapid muscle contractions and thus explains the absence of strong sexually dimorphic fiber type composition. Alternatively, selection for a phenotypic characteristic in one sex may also lead to expression of this feature in both sexes (Lande 1980).

The large inter-individual and interspecific variation in fiber type composition makes it difficult to interpret some of the more subtle differences between species. For example in Brown-headed Cowbird fast fibers appear to account for the larger than expected syringeal cross-sectional area, whereas in male Zebra Finch the high superfast fiber content does not lead to an exceptionally large cross-sectional area (Figure 7). Furthermore, there is no simple relationship between vocal repertoire size and fiber type composition. One could speculate that a large song repertoire containing rapid frequency and amplitude modulation, such as in the European Starling, would not only require very large syringeal muscles but also muscles that are most adapted for speed. However, this is not the case. While the syrinx musculature in European Starling is within the general pattern expected from the body size relationships, it falls distinctly off these relationships in other species. These trends indicate that more detailed studies of the relationship between acoustic behavior and syringeal muscle phenotype are likely to reveal how fiber type composition is finely tuned to the specific needs dictated by the unique acoustic features found in each vocal repertoire.

In conclusion, singing behavior or lack thereof in some female songbirds likely does not drive the strong sexual dimorphism in the superfast fiber composition of the syringeal muscles in these 2 species, because fiber type composition is similar in other species in which females are not known to sing. The differences in muscle fiber type composition between species and sexes may instead reflect muscle use for generation of the entire vocal repertoire, including song and calls.