Development of the hyolaryngeal architecture in horseshoe bats: Insights into the evolution of the pulse generation for laryngeal echolocation

Abstract

Background
The hyolaryngeal apparatus generates biosonar pulses in the laryngeally echolocating bats. The cartilage and muscles comprising the hyolarynx of laryngeally echolocating bats are morphologically modified compared to those of non-bat mammals, as represented by the hypertrophied intrinsic laryngeal muscle. Despite its crucial contribution to laryngeal echolocation, how the development of the hyolarynx in bats differs from that of other mammals is poorly documented. The genus Rhinolophus is one of the most sophisticated laryngeal echolocators, with the highest pulse frequency in bats. The present study provides the first detailed description of the three-dimensional anatomy and development of the skeleton, cartilage, muscle, and innervation patterns of the hyolaryngeal apparatus in two species of rhinolophid bats using micro-computed tomography images and serial tissue sections and compares them with those of laboratory mice. Furthermore, we measured the peak frequency of the echolocation pulse in active juvenile and adult individuals to correspond to echolocation pulses with hyolaryngeal morphology at each postnatal stage.

Results
We found that the sagittal crests of the cricoid cartilage separated the dorsal cricoarytenoid muscle in horseshoe bats, indicating that this unique morphology may be required to reinforce the repeated closure movement of the glottis during biosonar pulse emission. We also found that the cricothyroid muscle is ventrally hypertrophied throughout ontogeny, and that the cranial laryngeal nerve has a novel branch supplying the hypertrophied region of this muscle. Our bioacoustic analyses revealed that the peak frequency shows negative allometry against skull growth, and that the volumetric growth of all laryngeal cartilages is correlated with the pulse peak frequency.

Conclusions
The unique patterns of muscle and innervation revealed in this study appear to have been obtained concomitantly with the acquisition of tracheal chambers in rhinolophids and hipposiderids, improving sound intensity during laryngeal echolocation. In addition, significant protrusion of the sagittal crest of the cricoid cartilage and the separated dorsal cricoarytenoid muscle may contribute to the sophisticated biosonar in this laryngeally echolocating lineage. Furthermore, our bioacoustic data suggested that the mineralization of these cartilages underpins the ontogeny of echolocation pulse generation. The results of the present study provide crucial insights into how the anatomy and development of the hyolaryngeal apparatus shape the acoustic diversity in bats.

Full text

Development of the hyolaryngeal architecture in
horseshoe bats: Insights into the evolution of the
pulse generation for laryngeal echolocation
Taro Nojiri (  nojiri0805@gmail.com )
Juntendo University
Masaki Takechi
Juntendo University
Toshiko Furutera
Juntendo University
Nicolas L.M. Brualla
City University of Hong Kong
Sachiko Iseki
Tokyo Medical and Dental University
Dai Fukui
The University of Tokyo Fuji Iyashinomori Woodland Study Center, The University of Tokyo
Vuong Tan Tu
Vietnam Academy of Science and Technology
Fumiya Meguro
University of Tsukuba
Daisuke Koyabu
Tokyo Medical and Dental University
Research Article
Keywords: bats, laryngeal echolocation, hyolarynx, trachea, bioacoustics
Posted Date: September 6th, 2023
DOI: https://doi.org/10.21203/rs.3.rs-3325715/v1
License:   This work is licensed under a Creative Commons Attribution 4.0 International License. 
Read Full License
Additional Declarations: No competing interests reported.

Title: Development of the hyolaryngeal architecture in horseshoe bats: Insights into the evolution 1
of the pulse generation for laryngeal echolocation. 2
3
Taro Nojiri1*, Masaki Takechi1,2, Toshiko Furutera1,2, Nicolas L.M. Brualla3, Sachiko Iseki2, Dai 4
Fukui4, Vuong Tan Tu5,6, Fumiya Meguro7, Daisuke Koyabu2,3,7*. 5
6
1Graduate School of Medicine, Juntendo University, 2-1-1, Hongo, Bunkyo-ku, Tokyo 113-8421, 7
Japan 8
2Department of Molecular Craniofacial Embryology, Tokyo Medical and Dental University, 1-5-9
45 Yushima, Bunkyo-ku, Tokyo 113-8549, Japan 10
3Department of Infectious Diseases and Public Health, Jockey Club College of Veterinary 11
Medicine and Life Sciences, City University of Hong Kong, Hong Kong SAR, China 12
4The University of Tokyo Fuji Iyashinomori Woodland Study Center, Graduate School of 13
Agricultural and Life Sciences, The University of Tokyo, 341-2, Yamanaka, Yamanakako, 14
Yamanashi 401-05013, Japan 15
5Institute of Ecology and Biological Resources, Vietnam Academy of Science and Technology, 16
No. 18, Hoang Quoc Viet road, Cau Giay district, Hanoi, Vietnam 17
6Graduate University of Science and Technology, Vietnam Academy of Science and Technology, 18
No. 18, Hoang Quoc Viet road, Cau Giay district, Hanoi, Vietnam 19
7Research and Development Center for Precision Medicine, University of Tsukuba, 1-2 Kasuga, 20
Tsukuba-shi, Ibaraki 305-8550, Japan 21
22
Short running title: Hyolaryngeal development of horseshoe bats 23
24
*Corresponding author 25
Taro Nojiri: nojiri0805@gmail.com
26
Daisuke Koyabu: dsk8evoluxion@gmail.com 27
28

Abstract 29
Background 30
The hyolaryngeal apparatus generates biosonar pulses in the laryngeally echolocating bats. The 31
cartilage and muscles comprising the hyolarynx of laryngeally echolocating bats are 32
morphologically modified compared to those of non-bat mammals, as represented by the 33
hypertrophied intrinsic laryngeal muscle. Despite its crucial contribution to laryngeal 34
echolocation, how the development of the hyolarynx in bats differs from that of other mammals 35
is poorly documented. The genus Rhinolophus is one of the most sophisticated laryngeal 36
echolocators, with the highest pulse frequency in bats. The present study provides the first detailed 37
description of the three-dimensional anatomy and development of the skeleton, cartilage, muscle, 38
and innervation patterns of the hyolaryngeal apparatus in two species of rhinolophid bats using 39
micro-computed tomography images and serial tissue sections and compares them with those of 40
laboratory mice. Furthermore, we measured the peak frequency of the echolocation pulse in active 41
juvenile and adult individuals to correspond to echolocation pulses with hyolaryngeal 42
morphology at each postnatal stage. 43
Results 44
We found that the sagittal crests of the cricoid cartilage separated the dorsal cricoarytenoid muscle 45
in horseshoe bats, indicating that this unique morphology may be required to reinforce the 46
repeated closure movement of the glottis during biosonar pulse emission. We also found that the 47
cricothyroid muscle is ventrally hypertrophied throughout ontogeny, and that the cranial laryngeal 48
nerve has a novel branch supplying the hypertrophied region of this muscle. Our bioacoustic 49
analyses revealed that the peak frequency shows negative allometry against skull growth, and that 50
the volumetric growth of all laryngeal cartilages is correlated with the pulse peak frequency. 51
Conclusions 52
The unique patterns of muscle and innervation revealed in this study appear to have been obtained 53

concomitantly with the acquisition of tracheal chambers in rhinolophids and hipposiderids, 54
improving sound intensity during laryngeal echolocation. In addition, significant protrusion of 55
the sagittal crest of the cricoid cartilage and the separated dorsal cricoarytenoid muscle may 56
contribute to the sophisticated biosonar in this laryngeally echolocating lineage. Furthermore, our 57
bioacoustic data suggested that the mineralization of these cartilages underpins the ontogeny of 58
echolocation pulse generation. The results of the present study provide crucial insights into how 59
the anatomy and development of the hyolaryngeal apparatus shape the acoustic diversity in bats. 60
61
Keywords 62
bats, laryngeal echolocation, hyolarynx, trachea, bioacoustics 63

Background 64
Vocalization in mammals shows extraordinary acoustic and functional diversity, leading to 65
remarkable diversity in life histories. Among mammals, the mode of sound production varies with 66
feeding strategy [1], habitat [2], and social behavior [3]. The enormous diversity of sound 67
peculiarities such as frequency, intensity, and duration has been attributed to morphological 68
changes in the sound-producing organ, the hyolaryngeal apparatus, as reported in koalas [4], 69
ungulates [5], primates [6–12], manatees [13], rodents [14–18], and bats [19,20]. Structural 70
simplification of the larynx and laryngeal descent results in vocal complexity in humans [12]. 71
This evidence provides strong support for the idea that variations in mammalian sound signals are 72
shaped by drastic changes in hyolaryngeal anatomy. 73
Bats, which generate a biosonar sound with vocal folds, laryngeal cartilage, and intrinsic 74
laryngeal muscles for foraging and navigation (laryngeal echolocation), are ideal to explore the 75
link between hyolaryngeal anatomy and bioacoustics. Phylogenetically, extant bats comprise two 76
groups: Yinpterochiroptera (yinpterochiropterans) and Yangochiroptera (yangochiropterans) [21]. 77
Yinpterochiroptera comprises two lineages: Rhinolophoidea (rhinolophoids) and Pteropodidae 78
(pteropodids). Among these, rhinolophoids and yangochiropterans employ laryngeal echolocation 79
[22]. No pteropodids are capable of laryngeal echolocation, but a handful of pteropodid species 80
are known to employ non-laryngeal echolocation using tongue clicks [23] or wing flapping [24]. 81
The yangochiropterans emit biosonar sounds through the oral cavity, except for Phyllostomidae 82
(phyllostomids) and Nycteridae (nycterids), whereas the rhinolophoids, phyllostomids, and 83
nycterids emit biosonar sounds through the nostril by adjusting the acoustic properties of the nose 84
leaf at the snout [25]. Rhinolophidae (rhinolophids) and Hipposideridae (hipposiderids) species 85
in rhinolophoids and nycterids also possess two pairs of novel sound-producing organs, known as 86
tracheal chambers, at the lateral and dorsal sides of the tracheal rings [26]. This functions as a 87
Helmholtz resonating amplifier to improve sound intensity, possibly contributing to the emission 88

of echolocation pulses through the nostrils [26–28]. In contrast to yangochiropterans, 89
Rhinolophus and Hipposideros bats in rhinolophoids employ Doppler shift compensation (DSC), 90
which allows them to accurately localize targets by adjusting the sound frequency of ongoing 91
signals to fit them within the most suitable frequency for the primary auditory cortex [29]. For 92
this reason, the rhinolophids and hipposiderids are considered to possess one of the most 93
sophisticated echolocation systems among bats [22]. 94
Previous studies on the evolution of laryngeal echolocation have mostly focused on 95
auditory organs [30–39]. The cochlea, which transmits sound signals to the cerebral system in 96
vertebrates, is one of the most iconic organs because it is markedly enlarged in laryngeally 97
echolocating bats [30]. Additionally, laryngeally echolocating bats possess a fully ossified 98
stylohyal attached to the ectotympanic bone, which improves their sensitivity to ongoing signals 99
[32,40]. The morphology and development of these features have been well documented in the 100
context of the evolutionary origins of laryngeal echolocation in bats [31,32,35–38,41]. The 101
enlarged inner ear and ectotympanic-stylohyal unit develop heterotopically and heterochronically 102
in rhinolophoids and yangochiropterans, supporting the multiple-origin hypothesis of laryngeal 103
echolocation in bats [37]. The auditory organs for perceiving echolocation pulses have undergone 104
independent developmental changes, presumably reflecting the differences in biosonar pulse 105
emissions among bats. 106
Concerning the sound-producing organs of bats, a few studies have reported the 107
musculoskeletal morphology of the hyolaryngeal apparatus in adults [26,40,42–48]. Laryngeally 108
echolocating bats possess hypertrophied intrinsic laryngeal muscles supported by reinforced 109
cricoid, thyroid, and arytenoid cartilage [26,43,49]. In particular, the cricothyroid muscle is a 110
superfast muscle essential for the generation of high-frequency sounds in yangochiropterans [19]. 111
Given that the laryngeal cartilage associated with these muscles is gradually mineralized after 112
birth [28,45,47], it is predicted that the ontogeny of biosonar pulse emission is linked to laryngeal 113

growth [46]. However, the paucity of our knowledge on the development of the entire anatomy 114
of the hyolaryngeal apparatus and the acoustic peculiarities of the echolocation pulse restrict our 115
understanding of how morphological changes in the vocal apparatus shape the echolocation pulse 116
during ontogeny. 117
Therefore, we conducted the first detailed description of the pre- and postnatal 118
development of the hyolaryngeal apparatus and compared morphological development and pulse 119
ontogeny in horseshoe bats (the genus Rhinolophus). The hyolarynx was visualized three-120
dimensionally using microcomputed tomography (micro-CT). In addition, using a diffusible 121
iodine-based contrast-enhanced micro-CT technique [50,51], we described the morphology of 122
laryngeal cartilage and intrinsic laryngeal muscles in detail for the first time in bats. Using 123
immunohistochemistry, we described the three-dimensional morphology of the cartilage, muscles, 124
and the innervation patterns of the hyolaryngeal complex. Furthermore, we measured the 125
maximum frequency of each harmonic pulse, CF1, and CF2, from postnatal day 0 to adult 126
individuals and compared pulse ontogeny with morphological development. Our study highlights 127
the unique anatomical and developmental patterns of the laryngeal cartilage, cricothyroid muscle, 128
and cranial laryngeal nerve in laryngeally echolocating bats, providing tangible insights into how 129
the anatomy and development of the hyolarynx shape the acoustic diversity in bats. 130

Results 131
132
General morphology of the hyolaryngeal complex in horseshoe bats 133
134
Hyoid 135
The hyoid bone of horseshoe bats consists of five elements: basihyal, thyrohyal, ceratohyal, 136
epihyal, and stylohyal, whereas in mice, it includes only basihyal, epihyal, and stylohyal (Fig. 1). 137
In bats, the basihyal and thyrohyal bones were fully fused, and the boundary was not evident (Fig. 138
1A). The epihyal and stylohyal bones were elongated compared with those in mice. There was a 139
large gap between the basihyal bone and thyroid cartilage, unlike in mice (Fig. 1D, H). The 140
thyrohyoid bone of mice is connected to the thyroid cartilage only by the thyrohyoid muscle 141
attached to the ventral side of the basihyal and rostrolateral portion of the thyroid cartilage; 142
however, in horseshoe bats, the thyrohyoid muscle and caudal tip of the thyrohyal bone were 143
connected to the lateral side and thyroid cranial cornua, respectively (Fig. 1A). 144
145
Larynx 146
The cricoid cartilage had a sagittal crest in horseshoe bats, completely dividing the dorsal 147
cricoarytenoid muscle into left and right components (Fig. 1A-C). The thyroid cartilage had the 148
cranial and caudal cornua connected to the thyrohyal bone and muscular wings of the cricoid 149
cartilage in bats. The rostroventral tip of the thyroid cartilage was separated to form a pair of 150
protuberances, to which the thyroarytenoid muscle was attached (Fig. 1B). While the thyroid 151
cartilage of mice had one pair of the foramen thyroideum at the lamina of the thyroid cartilage, 152
through which the cranial laryngeal nerve passes, that of the thyroid cartilage was not opened in 153
the rhinolophoids. The cricothyroid muscle surrounding the cranial arch of the cricoid cartilage 154
was significantly enlarged in bats compared to that in mice. While the cricothyroid muscle was 155

separated into left and right components in mice, those of horseshoe bats met at the median line, 156
forming one large muscle complex. 157
158
Trachea 159
A pair of lateral and dorsal tracheal chambers were present in the ventral portion of the cranial 160
arch of the cricoid cartilage (Fig. 1A-D). The lateral tracheal chamber was completely attached 161
to the ventral side of the cricoid cartilage and the dorsal tracheal chamber. Both tracheal chambers 162
had a pouch-shaped structure surrounded by the glottis and tracheal rings and had internal cavities 163
connected to the respiratory tract. The ventral portion of the cricothyroid muscle was attached to 164
the dorsal side of the lateral tracheal chamber, denoting that the contraction of the tracheal 165
chamber was regulated by the common muscle to generate the biosonar pulse. 166
167
Innervation 168
Figure 2 shows the innervation pattern of the laryngeal muscles in CS22 of the horseshoe bat and 169
E18.5 of the laboratory mouse, which were three-dimensionally reconstructed using 170
immunohistochemical staining. The vagus nerve (cranial nerve X) has two branches, the cranial 171
and recurrent laryngeal nerves. In mice, the cranial laryngeal nerve passes through the foramen 172
thyroideum (Fig. 2E, F), whereas in bats, it passes below the thyroid cartilage and supplies the 173
cricothyroid muscle (Fig. 2B). In the cricothyroid muscle of bats, the cranial laryngeal nerve has 174
two branches, rostral and ventral (Fig. 2D). These two branches passed through the rostral and 175
ventral cricothyroid muscles, indicating that the ventral branch of the cranial laryngeal nerve 176
supplied contractions to the lateral tracheal chamber. The recurrent laryngeal nerve, previously 177
supplied in the postcranial direction, looped under the aortic arch, and reached the dorsal 178
cricoarytenoid muscle in both bats and mice (Fig. 2D, H). The recurrent laryngeal nerve passed 179
the dorsal cricoarytenoid muscle, was supplied to the lateral cricoarytenoid muscle, and finally to 180

the thyroarytenoid muscle (Fig. 2C, G). The supply pattern of the recurrent laryngeal nerve was 181
common between bats and mice. 182
183
Prenatal development of the hyolaryngeal complex in horseshoe bats and in mice 184
185
Hyoid development 186
Cartilaginous condensation began at CS17 (Fig. 3A, B) in Rhinolophus bats. The cranial cornu, 187
consisting of the ceratohyal, epihyal, and stylohyal, was separated from the hyoid body (basihyal) 188
and caudal cornu (thyrohyal), but was not segmented to form each component. The thyrohyoid 189
muscle, connecting the hyoid apparatus to the laryngeal cartilage, was present in the rostrolateral 190
portion of the hyoid complex and thyroid cartilage. 191
The condensed hyoid chondrocytes differentiated into ceratohyal, epihyal, and stylohyal 192
at CS20.5 (Fig. 3C, D). Periosteum was observed around the stylohyal bone, denoting that the 193
ossification started at this stage. The general morphology of the hyoid cartilage and muscle did 194
not change from CS20.5 onwards (Fig. 3E, F). 195
In Mus musculus, the epihyal, basihyal, thyrohyal, stylohyal cartilages, and thyrohyoid 196
muscles were already evident at E14.5, but the boundaries in each hyoid component were not 197
evident (Fig. 3G, H). A connection between the hyoid and larynx was also achieved at this stage, 198
after which there was no drastic change in the overall morphology of the hyolaryngeal complex 199
(Fig. 3K, L). 200
201
Laryngeal development 202
Chondrification of the cricoid, thyroid, and arytenoid began at CS17 in horseshoe bats. The cranial 203
arch was formed at the ventral margin of the cricoid cartilage that surrounds the glottis. The 204
sagittal crest projected dorsally from the medial portion of the cricoid cartilage. The cranial cornu 205

of the thyroid cartilage was dorsally projected to connect to the hyoid complex. The corniculate 206
processes of the arytenoid cartilage were by now in contact with one another. The cricothyroid, 207
cricoarytenoid, and thyroarytenoid muscles were present in CS17 (Fig. 4A-C). The cricothyroid 208
muscle was formed at the rostroventral portion of the thyroid cartilage and the rostrodorsal portion 209
of the cricoid cartilage, covering the cranial arch of the cricoid cartilage. The thyroarytenoid 210
muscle was formed at the ventral part of the arytenoid cartilage, protruding toward the thyroid 211
cartilage. The lateral and dorsal cricoarytenoid muscles were formed at the rostral portion of the 212
arytenoid cartilage and the caudal surface of the cricoid cartilage, respectively. The dorsal 213
cricoarytenoid muscle was separated into left and right components by the sagittal crest of the 214
cricoid cartilage (Fig. 4C). 215
The muscular wings of the cricoid cartilage projected bilaterally to form a joint with the 216
caudal cornu of the thyroid cartilage in CS20.5 (Fig. 4E, F). The sagittal crest of the cricoid 217
cartilage was projected to be greater than that of the CS17. The dorsal tip of the cranial cornua of 218
the thyroid cartilage was slightly curved. In dorsal view, the rostral tip of the thyroid cartilage was 219
curved, forming a V-shaped morphology (Fig. 4E). The vocal processes of the arytenoid cartilage 220
were projected medially around the glottis. The muscular process of the arytenoid cartilage also 221
projected bilaterally. The cricothyroid muscle further expanded toward the tracheal region to 222
surround the cranial arch of the cricoid cartilage. The thyroarytenoid muscle extended rostrally 223
and attached to thyroid cartilage. The dorsal cricoarytenoid muscle expanded dorsoventrally along 224
the sagittal crest of cricoid cartilage. 225
At CS22, the caudal cornu of the thyroid cartilage and muscular wings of the cricoid 226
cartilage projected in the ventral and bilateral directions, respectively. The projection of the front 227
portion occurred in the rostrolateral portion of the cranial arch of the cricoid cartilage. The rostral 228
tip of the thyroid cartilage became more convex than in the previous stage. The muscular 229
processes of the arytenoid cartilage protruded bilaterally to expand the attachment area of the 230

dorsal cricoarytenoid muscle. The cricothyroid muscle expanded medially and the cranial arch of 231
the cricoid cartilage was fully covered. The lateral portion of the cricothyroid muscle extended 232
beyond the front plate of the cricoid cartilage and reached the trachea. The thyroarytenoid muscle 233
extended rostrolaterally to expand the area attached to the thyroid cartilage. The dorsal 234
cricoarytenoid muscle extended dorsorostrally and was attached to the depression between the 235
corniculate and the vocal processes of the arytenoid cartilage. 236
Similarly, in mice, the cricoid, thyroid, and arytenoid cartilages were already 237
chondrified at E14.5 (Fig. 4J). The sagittal crest of the cricoid cartilage was slightly projected. 238
The cranial arch of the cricoid cartilage surrounding the glottis was evident. The thyroid cranial 239
cornu projected dorsally to form a joint with the hyoid complex. The rostral tip of the thyroid 240
cartilage was convex in the dorsal view. The arytenoid cartilage had already undergone a 241
corniculate process. The cricothyroid muscle was also evident, running between the ventrocaudal 242
and dorsorostral portions of the thyroid cartilage. The thyroarytenoid muscle was formed at the 243
rostral site of the arytenoid cartilage and attached to the caudal side of the thyroid cartilage. The 244
lateral cricoarytenoid muscle at the rostroventral site of the arytenoid cartilage remained small. 245
The dorsal cricoarytenoid muscle was still not in contact with the lateral cricoarytenoid muscle 246
and was not separated by the sagittal crest of the cricoid cartilage, similar to that in horseshoe bats 247
(Fig. 4L). At E16.5, the cricoid cartilage expanded ventrally to expand the attachment area of the 248
dorsal cricoarytenoid muscle. The thyroid cartilage was in contact with the muscular wings of the 249
cricoid cartilage. In contrast to horseshoe bats, a pair of the foramen thyroideum was opened at 250
the lateral lamina of the thyroid cartilage, through which the cranial laryngeal nerve passes. From 251
E16.5 onwards, the morphology of the laryngeal cartilage did not change drastically (Fig. 4M-O). 252
At E18.5, the dorsal cricoarytenoid muscle was separated by the caudally projected sagittal crest 253
of the cricoid cartilage but was not fully divided, as in bats (Fig. 4P-R). As for the cricothyroid 254
muscle, the rostral tips of the left and right muscles did not meet or reach the trachea; thus, the 255

cranial arch of the cricoid cartilage remained exposed, unlike in horseshoe bats. 256
257
Tracheal development 258
In horseshoe bats, chondrification of the tracheal ring was not evident at CS17, whereas 259
chondrocytes were condensed in the region of the lateral tracheal chamber, denoting that the 260
lateral tracheal chamber is a novel laryngeal cartilage acquired in horseshoe bats (Fig. 5A). 261
Condensed chondrocytes in the lateral tracheal chamber were continuous with those in the cricoid 262
cartilage (Fig. 5A-2). At CS22, the tracheal rings began to chondrify (Fig. 5B). Simultaneously, 263
the dorsal tracheal chamber was chondrified ventrally into the lateral tracheal chamber. The 264
condensed chondrocytes in both the lateral and dorsal tracheal chambers were separated to form 265
the left and right tracheal chambers (Fig. 5B-2, B-3). The lateral tracheal chamber covered the 266
tracheal region from the cranial arch of the cricoid cartilage to the third tracheal ring, while the 267
dorsal tracheal chamber covered the fourth to sixth tracheal rings. The lateral tracheal chamber 268
was articulated with the cranial arch of the cricoid cartilage (Fig. 5B-2), whereas the dorsal 269
tracheal chamber was connected to the sixth tracheal ring (Fig. 5B-4). The ventral portion of the 270
cricothyroid muscle reached the caudal side of a pair of lateral tracheal chambers. 271
In laboratory mice, tracheal rings were not evident at E14.5 with all components of the 272
laryngeal cartilage (Fig. 5C). Tracheal rings were chondrified ventrally to the cranial arch of the 273
cricoid cartilage at E16.5 (Fig. 5D), after which there was no drastic shift in the morphology of 274
the tracheal region (Fig. 5E). 275
276
Pre- and postnatal ossification and mineralization of the hyolaryngeal apparatus 277
The presence of ossification of all craniocervical bony elements at each stage is summarized in 278
Table S1-3. In horseshoe bats, ossification of the stylohyal occurred at CS19 and other hyoid 279
components did not start until CS21 (Fig. 6). The epihyal and thyrohyal were simultaneously 280

ossified in the rostral region of the stylohyal in the CS22. In CS24 mice, the basihyal and 281
ceratohyal regions ossified between the left and right thyrohyal and rostral regions of the epihyal 282
region, respectively. Simultaneously, mineralization of the laryngeal cartilaginous elements, 283
cricoid, thyroid, and arytenoid, became evident. The center of mineralization of the cricoid 284
cartilage was observed in the proximal part of the cranial arch. The mineralization of the thyroid 285
cartilage starts from the lateral part, at which point the future joint with the cricoid cartilage is 286
formed. Mineralization of the arytenoid cartilage begins at the caudal tip of the corniculate process. 287
In postnatal day 0 (P0) specimens of horseshoe bats, the mineralized cricoid cartilage further 288
expanded to form muscular wings and a cricoid lamina with the sagittal crest. The proximal part 289
of the cranial arch extended rostrally. The muscular wings of the cricoid cartilage were in contact 290
with those of the thyroid cartilage. The small ossification center of the arytenoid cartilage in CS24 291
expanded to form corniculate and muscular processes immediately after birth. At P14, the 292
stylohyal, epihyal, ceratohyal, and thyrohyal regions extend caudally. The thyrohyal and basihyal 293
were fully fused and the suture was absent. The sagittal crest of the cricoid cartilage extended 294
caudally. The cranial arch of the cricoid cartilage was further mineralized to fully cover the glottis. 295
A pair of cranial cornua of the thyroid cartilage was mineralized caudally to the basihyal. The 296
vocal process projected slightly at the ventral base of the arytenoid cartilage. The left and right 297
dorsal tips of the corniculate processes were in contact with each other. From this stage onwards, 298
mineralization of the lateral tracheal chamber began. The lateral tracheal chamber was in contact 299
with the ventral side of the cricoid cartilage, forming an internal cavity connected to the 300
respiratory tract. The sagittal crest of the cricoid cartilage extended dorsally at P21. The left and 301
right mineralized cranial cornua of the thyroid cartilage were merged. The left and right vocal 302
processes of the arytenoid cartilage were also in contact with the median line. From P21 onwards, 303
the dorsal tracheal chamber began to mineralize. The mineralized portion of the dorsal tracheal 304
chamber was in contact with the ventral portion of the lateral tracheal chamber. In adult horseshoe 305

bats, the sagittal crest of the cricoid cartilage was sharper and thicker than that in juveniles. The 306
mineralized lateral and dorsal tracheal chambers also thickened until sexual maturation. 307
Figure 7 shows the comparative anatomy of postnatal hyoid ossification and laryngeal 308
mineralization in bats and mice. The relative ossification timing of the hyoid components is shown 309
in Fig. 8 and Table S3. While only the basihyal was ossified in P0 mice, all components of the 310
hyoid apparatus started their ossification just after birth in bats. Indeed, except for the basihyal, 311
almost all components of the hyoid apparatus ossified at an earlier stage in bats than in mice. 312
Laryngeal mineralization was limited to the cricoid and thyroid cartilages in mice, whereas in 313
bats, three laryngeal cartilages–the cricoid, thyroid, and arytenoid cartilages–were mineralized. 314
Mineralization of the thyroid cartilage occurred in almost all parts of the mice, whereas in bats, 315
the mineralized area of the thyroid cartilage was partial (Fig. 7A-F). Regarding the mineralization 316
rate, laryngeal mineralization occurred in a relatively short period (P14-adult), while that of bats 317
proceeded over the long term (CS24-adult). 318
319
Allometric growth of the larynx and echolocation pulse ontogeny 320
The basic statistics and results of the reduced major axis (RMA) regression analyses of 321
mineralized laryngeal cartilage against the cubed geometric mean (GM3) in horseshoe bats are 322
shown in Table 3 and Fig. 9A-D. A significant correlation with log10-transformed GM3 was 323
detected in all log10-transformed volumes of mineralized laryngeal cartilage: cricoid cartilage (r 324
= 0.90, p < .001), thyroid cartilage (r = 0.88, p < .001), arytenoid cartilage (r = 0.93, p < .001), 325
and lateral tracheal chamber (r = 0.88, p < .001). The allometric exponents of all laryngeal 326
cartilages were above 1.0 (i.e. positive allometry) as follows: 2.01 in the cricoid cartilage (95% 327
CI from 1.21 to 2.83), 3.20 in the thyroid cartilage (95% CI from 1.82 to 4.39), 3.93 in the 328
arytenoid cartilage (95% CI from 2.72 to 4.88), and 5.66 in the lateral tracheal chamber (95% CI 329
from 3.14 to 7.69). The y-intercept values were as follows: -5.68 in the cricoid (95% CI from -330

7.95 to -3.43), -9.46 in the thyroid (95% CI from -12.82 to -5.59), -11.68 in the arytenoid (95% 331
CI from -14.36 to -8.23), and -16.47 in the lateral tracheal chamber (95% CI from -22.21 to -9.36). 332
The results of the RMA regression analyses of the log10-transformed maximum 333
frequency of the first-harmonic (CF1) and second-harmonic (CF2) pulses against the log10-334
transformed GM3 are shown in Fig. 9E-F. Significant correlations with the GM were detected for 335
CF1 (r = 0.76, p < .05) and CF2 (r = 0.76, p < .05). The regression slope of both CF1 and CF2 336
significantly shows negative allometry as follows: 0.47 in CF1 (95% CI from 0.04 to 0.73) and 337
0.46 in CF2 (95% CI from 0.05 to 0.73). The y-intercept value was 0.39 in the CF1 (95% CI from 338
-0.36 to 1.61) and 0.70 in the CF2 (95% CI from –0.04 to 1.88). 339
Figure 10 shows the ontogeny of the sonograms of the CF1 and CF2 pulses. The 340
echolocation pulse was not confirmed until P13; thus, the biosonar pulse emission was acquired 341
for at least two weeks after birth. At P14, the maximum frequencies of CF1 and CF2 were 49.58–342
49.99 kHz and 99.20–99.98 kHz, respectively. The values of CF1 and CF2 in P21 were 50.67 kHz 343
and 101.23 kHz, respectively. In P30, the obtained values of each harmonic pulse were 50.89–344
51.28 kHz and 101.83–102.67 kHz. The maximum value of each harmonic pulse rose to 51.69 345
kHz and 103.37 kHz in P45. In adult individuals, the maximum frequencies of CF1 and CF2 finally 346
reached 54.14 kHz and 108.25 kHz. 347
348
Relationship between laryngeal mineralization and pulse ontogeny 349
Table 4 and Fig. 11 illustrate the results of generalized linear model (GLM) analyses in 350
horseshoe bats to investigate the relationship between the log10-transformed volume of the 351
mineralized laryngeal cartilage and the log10-transformed peak frequencies of the CF1 and CF2 352
components of the echolocation pulse. Significant positive associations with the maximum 353
frequency of the CF1 component were observed in the cricoid (slope = 0.21, standard error [SE] 354
= 0.04, p < .001), thyroid (slope = 0.15, SE = 0.06, p < .001), arytenoid (slope = 0.15, SE = 0.04, 355

p < .001), and lateral tracheal chambers (slope = 0.09, SE = 0.02, p < .001). Regarding the 356
associations with the maximum frequency of the CF2 component, significant positive associations 357
were detected in the cricoid cartilage (slope = 0.21, SE = 0.04, p < .001), thyroid cartilage (slope 358
= 0.15, SE = 0.06, p < .001), arytenoid cartilage (slope = 0.15, SE = 0.04, p < .001), and lateral 359
tracheal chamber (slope = 0.08, SE = 0.02, p < .001). 360
361

Discussion 362
363
General morphology of the hyolaryngeal complex in the horseshoe bats 364
365
The hyolaryngeal morphology of horseshoe bats has the following six unique characteristics 366
compared to that of mice: (1) a spatial gap between the hyoid complex and thyroid cartilage; (2) 367
enlarged intrinsic muscles (cricothyroid, thyroarytenoid, dorsal, and lateral cricoarytenoid 368
muscles); (3) sagittal crest of the cricoid cartilage; (4) a fully separated dorsal cricoarytenoid 369
muscle; (5) the presence of lateral and dorsal tracheal chambers, and (6) the additional branch of 370
the cranial laryngeal nerve innervating the cricothyroid muscle. Characteristics (2), (3), and (5) 371
are consistent with the descriptions of Rhinolophus hipposideros [43] and Hipposideros caffer 372
[26]. As reported in other laryngeally echolocating bats [19,49], the intrinsic laryngeal muscles 373
of horseshoe bats are hypertrophied compared to those of mice, and the laryngeal cartilage is 374
drastically modified to provide attachment sites for these muscles. The overall pattern of 375
innervation of the intrinsic laryngeal muscles of horseshoe bats was identical to that of mice, 376
except for a novel branch of the cranial laryngeal nerve and the absence of the foramen 377
thyroideum (Fig. 2). 378
Our observations identified a hyoid-larynx gap, showing that the connection between 379
the hyoid and larynx is evident in horseshoe bats (Fig. 1D). This structure has also been confirmed 380
in other laryngeally-echolocating bats such as Desmodus rotundus [32] and Artibeus jamaicensis 381
[47]. Instead, the stylohyal bone is articulated with the ectotympanic bone, which contributes to 382
the direct conduction of ongoing echolocation signals in laryngeally echolocating bats [32] and 383
tree mice [16]. However, these gaps and articulations are absent in pteropodids (Giannini et al., 384
2006) and non-bat mammals (Fig. 1G). Therefore, in laryngeally echolocating bats, the hyoid 385
apparatus itself is likely to be extended rostrocaudally, providing a gap between the hyoid 386

apparatus and laryngeal cartilage and articulation between the stylohyal and ectotympanic. 387
Biosonar sound requires rotation of the thyroid cartilage to increase the tension of the vocal folds 388
with the superfast muscle [19]. Acoustic response analyses have revealed that the bony connection 389
between the stylohyal and ectotympanic transfers sound signals at the most susceptible amplitude 390
to the auditory bullae [40,48]. Thus, as a result of the release from the tight connection, the hyoid 391
complex and laryngeal cartilage might become mutually movable, providing higher flexibility for 392
bone conduction and thyroid rotation, respectively. 393
The dorsal cricoarytenoid muscle was completely divided by the projection of the 394
sagittal crest of the cricoid cartilage (Fig. 1C). In mice, the dorsal cricoarytenoid muscle was 395
attached to the cricoid and arytenoid cartilages as a single muscle (Fig. 1G), whereas in bats, the 396
separated left and right muscles were attached to the left and right arytenoid cartilages (Fig. 1C). 397
This indicates that the separated dorsal cricoarytenoid muscle contracts independently, 398
contributing to frequency adjustment in horseshoe bats. Generally, the dorsal cricoarytenoid 399
muscle adducts with the arytenoid cartilage and adjusts the tension of the vocal folds during 400
vocalization in mammals [52]. High-duty cycle (HDC) echolocating bats such as horseshoe bats 401
emit sound signals multiple times per echolocation [22]. Previous studies have reported that the 402
sagittal crest of the cricoid cartilage is also protruding in laryngeally echolocating lineages with 403
oral cavities (Mormoopidae [53] and Vespertilionidae [43,54]), whereas that of the non-404
laryngeally echolocating bats, the pteropodids, is relatively smooth [55]. In contrast to 405
rhinolophids (this study) and hipposiderids [26], the sagittal crests of the cricoid cartilages in 406
these groups does not fully divide the dorsal cricoarytenoid muscles, indicating that this muscle 407
pattern is unique to HDC-echolocating bats. Therefore, the independent contraction of the dorsal 408
cricoarytenoid muscle caused by the development of the sagittal crest of the cricoid cartilage 409
would increase the abduction efficiency of the arytenoid cartilage during HDC pulse emission in 410
horseshoe bats. 411

We confirmed that the cricothyroid muscle, possibly innervated by a novel branch of the 412
cranial laryngeal nerve, had expanded toward the lateral tracheal chamber (Fig. 2). The lateral 413
and dorsal tracheal chambers have been hypothesized to be key features involved in pulse 414
emission through the nostrils in the genera Rhinolophus and Hipposideros in rhinolophoids and 415
nycterids [26,43]. Tracheal chambers have been theorized to function as Helmholtz resonators 416
that increase sound intensity for laryngeal echolocation [26,27]. Because biosonar pulse 417
generation is underpinned by the superfast contraction of the cricothyroid muscle in laryngeally 418
echolocating bats (Elemans et al., 2011), the expansion and contraction of the tracheal chambers 419
are likely synchronized with that of the cricothyroid muscle. If the hypertrophied cricothyroid 420
muscle and additional branches of the cranial laryngeal nerve are required for acoustic function 421
in the tracheal chambers, this innervation pattern would be a synapomorphy of rhinolophid and 422
hipposiderid bats. Further investigations on the innervation patterns of nycterids that also have 423
tracheal chambers (Denny, 1976) are needed to test this hypothesis. 424
425
Prenatal development of the hyolaryngeal complex in horseshoe bats and in mice 426
We found that the hyolaryngeal complex reached adult morphology in bats and mice at the early 427
fetal stage (Figs. 3A and 4A). The bilaterally extended lesser and greater horns of the hyoid in 428
horseshoe bats became evident at an earlier stage than those in CS17. In humans, the second 429
pharyngeal cartilage forms the elongated lesser horn of the hyoid at Carnegie Stage 18, followed 430
by the development of the greater horn of the hyoid [56]. These cases suggest that the elongation 431
of each horn of the hyoid is due to chondrocyte proliferation during the early fetal stage. We also 432
confirmed that the thyrohyoid muscle increasingly develops toward the rostral side and merges at 433
the median line in mice but not in horseshoe bats (Fig. 3E, F). Around Carnegie Stage 20 in 434
humans, the thyrohyoid muscle connects the greater horn of the hyoid and the thyroid cartilage 435
[57], similar to rhinolophids. Together, the loosened hyoid-larynx connection of rhinolophids may 436

result from differences in the initial construction of the thyrohyoid muscle, at least until CS17. 437
Our observations confirmed that each intrinsic laryngeal muscle became distinguishable 438
at CS17 in bats and E14.5, in mice (Figs. 3 and 4). The time at which patterning of the intrinsic 439
laryngeal muscles occurs in mice is consistent with the observations in the mouse embryo [58]. 440
We also found that the left and right components of the dorsal cricoarytenoid muscle were present 441
at the boundary of the sagittal crest, similar to CS17 (Fig. 3A). In mice, the dorsal cricoarytenoid 442
muscle develops as a single muscle and is not divided by the sagittal crest throughout ontogeny 443
(Fig. 3E-L). Because this pattern has also been described in rat embryos [59], a single dorsal 444
cricoarytenoid muscle is likely to be conserved in rodents. In humans, the posterior cricoarytenoid 445
muscle develops from two independent muscles and maintains a separate morphology until 446
adulthood [60]. These facts suggest that the dorsal cricoarytenoid muscle of the horseshoe bat is 447
not secondarily separated by the projection of the sagittal crest, but originally has two 448
developmental centers. Despite less documented myogenesis in non-model organisms, separated 449
dorsal cricoarytenoid muscles in adults have been reported in elk, red deer [5], minipig [61], and 450
horses [62]. Therefore, it is possible that a single dorsal cricoarytenoid muscle in mice is the 451
derived phenotype in mammals, and the separated muscle in horseshoe bats is a synapomorphic 452
trait in laurasiatherians. 453
We observed that the rostral tip of the cranial arch of the cricoid cartilage projected to 454
split the cricothyroid muscle in horseshoe bats (Fig. 4D, G). The ventral component of the 455
cricothyroid muscle was attached to the ventral side of the cranial arch, and the muscular wings 456
of the cricoid cartilage were formed at CS20.5. Contrary to its development in mice, the ventral 457
part of the cricothyroid muscle appeared to be newly formed and merged with the dorsal part (Fig. 458
4D). Therefore, the hypertrophied cricothyroid muscle in horseshoe bats may not be established 459
by the expansion of a single component, but by the fusion of the original cricothyroid muscle and 460
novel muscle components. Considering that this muscle is also attached to the lateral tracheal 461

chamber, this pattern may be unique to rhinolophid and hipposiderid bats. The cricothyroid 462
muscle of laryngeally echolocating bats has been reported to be hypertrophied [26] and capable 463
of high-speed contraction in a few species (Myotis daubentonii [19] and Myotis lucifugus [63]). 464
Although the early development of the cricothyroid muscle has not yet been documented in other 465
laryngeally echolocating lineages, the present study implies that the hypertrophied cricothyroid 466
muscle is established through different developmental pathways among lineages with different 467
biosonar pulse generation capabilities. Moreover, combined with observations in postnatal 468
individuals (Fig. 1), the development of the cricothyroid muscle may proceed until birth. A similar 469
pattern has been reported in horseshoe petrosal, in which the size growth is unusually prolonged 470
until sexual maturation, leading to a remarkably enlarged cochlea in horseshoe bats [35]. The 471
present study implies that the development of multiple organs related to both the reception and 472
emission of biosonar sounds is prolonged concomitantly with the sophistication of laryngeal 473
echolocation in horseshoe bats. 474
475
On the embryonic origins of the tracheal chambers in horseshoe bats 476
We have described the early development of the lateral and dorsal tracheal chambers in horseshoe 477
bats (Fig. 5). Immunohistochemistry for the Sox9 protein identified that the lateral and dorsal 478
tracheal chambers developed by chondrocyte proliferation, such as the laryngeal cartilage (Fig. 479
5A-2). It has been hypothesized that tracheal chambers are formed by the modification of the 480
tracheal ring (Denny et al. 1976). Histological observations indicated that the condensed 481
chondrocyte mass comprising the lateral tracheal chamber originated from the ventral portion of 482
the cricoid cartilage, whereas the tracheal rings had not yet chondrified (Fig. 5A). Thus, we 483
suggest that the lateral and dorsal tracheal chambers are cricoid-derived laryngeal cartilages 484
acquired in rhinolophids and hipposiderids due to the expansion of the caudal tip of the cricoid 485
cartilage. It has been widely accepted that the rhinolophids and hipposiderids significantly radiate 486

their bioacoustics strategy, as represented by the emission of HDC pulses [22] and DSC, allowing 487
them to improve the accuracy of target localization [29]. Tracheal chambers are known to function 488
as pulse amplifiers for the reinforcement of sound intensity because of their susceptibility to 489
attenuation during emission through the nostrils [26,27]. In light of this, key innovations in the 490
cricoid cartilage might have led to the sophistication of laryngeal echolocation in tandem with the 491
acquisition of HDC and DSC capabilities. The concept of “key innovations” has been defined in 492
evolutionary developmental biology as a substantial change of the characters that have triggered 493
lineage diversification [64,65]. In accordance with this definition, tracheal chambers found 494
explicitly in laryngeally-echolocating bats emitting pulses with their nostrils (rhinolophids, 495
hipposiderids, and nycterids) should be regarded as key innovations in the larynx. 496
Genetic studies have reported that the development of respiratory organs such as the 497
larynx, trachea, and lungs is regulated by the expression of Fgf10 [66], chordin [67], Fuz, Gli3 498
[68], β-catenin [69], and Tbx1 [70]. In particular, it has been reported that Fgf10 knockout 499
experiments induce the disorganization of mesenchymal cells at the distal end of the trachea and 500
the absence of lung buds [66]. Given this, changes in the genetic factors represented by Fgf10 501
might result in changes in the distribution of mesenchymal cells, thereby inducing the expansion 502
of the cricoid cartilage in horseshoe bats. Further investigations using gene expression analysis 503
are required to uncover the genetic basis underlying the evolution of tracheal chambers. 504
505
Pre- and postnatal ossification of the hyolaryngeal apparatus 506
This study described hyoid ossification (Figs. 7 and 8) and estimated the relative ossification 507
timing of each element (Fig. 8). The timing of ossification of the hyoid components in horseshoe 508
bats was accelerated relative to that of mice, except for the basihyal component (Fig. 8), although 509
hyoid condensation was simultaneously observed at the early fetal stage in both species (Fig. 3). 510
This indicates that the ossification sequence was not necessarily consistent with the 511

chondrification sequence. In mice, the basihyal was the only hyoid component to ossify at birth 512
(Fig. 7G), whereas in horseshoe bats, all bony elements in the hyoid apparatus were already 513
ossified (Fig. 7A). In laryngeally echolocating bats, the stylohyal bone is articulated with an 514
ectotympanic [32], which possibly facilitates the transmission of the ongoing signal into the inner 515
ear through bone conduction [40]. A recent biomechanical analysis revealed that the closer the 516
bony element was to the basihyal bone, the higher the sound pressure is (Snipes and Carter, 2022). 517
The ongoing echolocation pulse, which is produced in the larynx, transfers to the thyrohyal and 518
basihyal regions and subsequently propagates into the ceratohyal, epihyal, stylohyal, and 519
ectotympanic regions. Our results suggest that construction of the stylohyal-ectotympanic unit is 520
preceded relatively by the thyrohyal-basihyal unit. 521
Our analyses revealed that the thyroid, cricoid, and arytenoid cartilages began those 522
mineralization immediately before birth in horseshoe bats. The cricoid and arytenoid cartilages 523
were entirely mineralized, whereas mineralization of the thyroid cartilage was limited to the 524
lateral and rostral parts, even in the adult stage (Fig. 6). The pattern of laryngeal mineralization 525
in horseshoe bats contrasts with that in phyllostomid bats, which possess non-mineralized thyroid 526
and arytenoid cartilages [26,40,45]. It has been suggested that laryngeal cartilage, which requires 527
dynamic motion, tends to be more mineralized than static cartilages such as the cricoid cartilage 528
[45]. Horseshoe bats produce high-intensity sounds in contrast to phyllostomids that emit low-529
intensity pulses [71]; thus, the highly mineralized cricoid and arytenoid cartilages in horseshoe 530
bats would provide physical support for the superfast contraction of the hypertrophied 531
cricothyroid muscle. Considering that the mineralized lateral and rostral parts of the thyroid 532
cartilage correspond to the attachment sites of the cricothyroid and thyroarytenoid muscles, 533
respectively (Fig. 1A-D), limited mineralization may be due to the high functional requirements 534
of these two muscles for laryngeal echolocation. The limited mineralization of the thyroid 535
cartilage, on the other hand, may be due to the high functional requirements of the cricothyroid 536

and thyroarytenoid muscles for pulse generation. Indeed, phyllostomid bats without mineralized 537
thyroid and arytenoid cartilage possess a relatively smaller cricothyroid muscle than other 538
echolocating bats [28,44]. Taken together, the differences in the mineralized laryngeal cartilage 539
between the rhinolophids and hipposiderids groups and phyllostomids may imply that these two 540
nasally-echolocating lineages employ different morphologies and kinetics of the laryngeal muscle 541
during laryngeal echolocation. 542
Mineralization of the lateral and dorsal tracheal chambers began at P14 and P21, 543
respectively (Figs. 6E, F, 7 B, C). Although the tracheal chambers do not have attachment sites 544
for the other muscles, except for the cricothyroid muscle, almost all the tracheal chambers were 545
mineralized, such as the cricoid and arytenoid cartilages. Chondrification was already evident at 546
CS17 (Fig. 5A), and connections between the lateral tracheal chamber and cricoid cartilage and 547
between the dorsal tracheal chamber and third tracheal ring were established until CS22 (Fig. 5B). 548
The air sac in the tracheal chambers prevents the echolocation pulse from being reflected from 549
the lungs towards the auditory organs [72]. The tracheal chambers would vibrate during the 550
emission of the biosonar pulse and are exposed to mechanical stress, thereby leading to 551
mineralization of the entire cartilage. Following the mineralization of the other laryngeal 552
cartilages, the structurally strengthened cartilages would increasingly reinforce the capability of 553
biosonar pulse emission. 554
555
Relationship between development and pulse ontogeny 556
We confirmed that no echolocation pulses were observed at the newborn (P0) stage (Fig. 10), as 557
reported in a previous study [73]. In the P0 specimen, the basihyal bone was not fused with the 558
thyrohyal bone, and each component of the lesser horn (ceratohyal, epihyal, and stylohyal) of the 559
hyoid was not in contact, indicating that the capability of bone conduction was still immature at 560
this stage (Figs. 6D, 7A). Furthermore, mineralization of the rostral part of the thyroid cartilage, 561

cranial arch of the cricoid cartilage, and lateral tracheal chamber was not observed at this stage. 562
Because the basihyal-thyrohyal unit is most susceptible to sound pressure during laryngeal 563
echolocation [40], the establishment of a strongly supported hyoid apparatus may be essential for 564
the acquisition of biosonar pulse emission. The echolocation pulse is known to gradually develop 565
from spontaneous calls used for mother-infant communication when juvenile individuals start 566
powered flights [74]. Our findings suggest that the progression of laryngeal mineralization 567
reinforces the rapid contraction of the intrinsic laryngeal muscles and consequently allows bat 568
pups to distinguish the echolocation pulse from precursor calls. As tracheal chambers improve the 569
sound intensity for laryngeal echolocation [26,27,72], well-mineralized tracheal chambers are 570
required for the maturation of high-intensity pulses in horseshoe bats [1]. Therefore, in contrast 571
to phyllostomids with low-intensity calls, the structural demand for the endurance of high-572
intensity signals is likely reflected in the progressive mineralization of the laryngeal cartilages of 573
horseshoe bats. 574
We detected positive allometric patterns in all laryngeal cartilages compared to the 575
cubed geometric mean (Fig. 9A-D). Regarding the cricoid cartilage, mineralization occurred 576
within a short period immediately before birth (Fig. 9A). The cricoid cartilage is surrounded by 577
the dorsal cricoarytenoid and cricothyroid muscles (Fig. 1B-D). We suggest that the precedence 578
of the mineralization of the cricoid cartilage allows horseshoe bats to promptly reinforce the 579
foundation of their superfast muscles. Moreover, the regression slope of the peak frequency of the 580
echolocation pulse against the cubed geometric mean showed a negative allometric pattern (Fig. 581
9E, F). The correlation between body mass and sound frequency in mammals has been widely 582
recognized as size-frequency allometry [75]. A previous study reported a significant positive 583
relationship between forearm length and the frequency of the second harmonic pulse in the 584
juvenile stage of R. cornutus [74] and R. ferrumequinum [76], but the negative allometric pattern 585
from juvenile to adult stages found in the present study (Fig. 9) indicates that the pulse peak 586

frequency does not necessarily increase at a rate similar to that of body size. 587
Our GLM analyses identified significant correlations between the cricoid thyroid, 588
arytenoid cartilage, and tracheal chambers and the peak frequencies of the CF1 and CF2 589
components (Fig. 11). Since the mineralized parts of the laryngeal cartilage correspond to the 590
attachment sites of the cricothyroid, thyroarytenoid, and cricoarytenoid muscles (Fig. 1A-D), 591
these four laryngeal cartilages are likely to be highly susceptible to mechanical stress during 592
muscle contraction. The generation of high-frequency sounds requires superfast muscle (i.e., the 593
cricothyroid muscle) control of the vocal cords [19]. Abduction of the glottis to vibrate the vocal 594
fold involves contraction of the dorsal cricoarytenoid muscle. It has been suggested that tracheal 595
chambers contribute to sound intensity [26,27,49,72]. Given that the ventral part of the 596
cricothyroid muscle is attached to the lateral tracheal chambers, mineralization of the tracheal 597
chambers can be promoted by superfast contraction of this muscle. Thus, the ontogeny of the 598
pulse peak frequency was also reflected in the mineralization of the tracheal chambers. Hence, 599
we suggest that the ontogeny of the peak frequency of the echolocation pulse is underpinned by 600
the mineralization of these four laryngeal cartilages in horseshoe bats. 601
602
Implications for the evolution of the laryngeal echolocation in bats 603
We found that horseshoe bats possess a unique morphology, such as hypertrophied intrinsic 604
laryngeal muscles (Fig. 1B) and a sagittal crest (Fig. 1C), as reported in previous studies [26,43]. 605
Regarding hyoid morphology, we identified that horseshoe bats possess the following five hyoid 606
elements: basihyal, thyrohyal, ceratohyal, epihyal, and stylohyal (Figs. 1A-D, 3A-C, 6, 7). A novel 607
pattern of innervation of the cranial laryngeal nerve was also identified using 608
immunohistochemical staining (Fig. 2). Considering that the novel branch of the cranial laryngeal 609
nerve runs into the ventrally expanded portion of the cricothyroid muscle, the hypertrophied 610
cricothyroid muscle comprises two components:1) homologous cricothyroid muscle in general 611

mammals and 2) novel muscle acquired along with the acquisition of the tracheal chambers in 612
horseshoe bats (Fig. 2). Our findings corroborate that the evolution of sophisticated biosonars, 613
such as DSC, in horseshoe bats might be underpinned by morphological changes in the 614
hyolaryngeal complex. 615
Hyoid morphology has been recognized as a key component of laryngeal echolocation 616
[31,32,37,40,41,44,49,53,77–80]. Regarding the entire morphology of the hyoid apparatus in bats, 617
it has been suggested that the number of hyoid components varies among lineages [44]. In contrast 618
to the hyoid apparatus with the segmented lesser cornu in horseshoe bats (Figs. 1A-D, 6, 7A-C), 619
for example, vespertilionid bats do not have the epihyal and ceratohyal, forming the lesser cornu 620
only with the stylohyal bone [81]. The genus Myotis, on the other hand, is known to have a lesser 621
cornu formed by the two bony elements [81], although the homology of each component in 622
horseshoe bats is still unclear. Because morphological differences in the hyoid apparatus are 623
related to variations in the efficiency of sound propagation [40], morphological changes in the 624
hyoid apparatus may contribute to the diversification of acoustic properties, such as sound 625
frequency, duration, and intensity, in laryngeally echolocating bats. 626
The present study identified unique characteristics such as a possibly novel branch of 627
the cranial laryngeal nerve in rhinolophids and hipposiderids. This lineage may have achieved 628
pulse generation with different laryngeal kinetics of the superfast muscle compared to that of 629
yangochiropterans [19]. Regarding pteropodids, the sister clade of the rhinolophoids within the 630
yinpterochiropterans, the morphology of their laryngeal muscles has been reported to be similar 631
to that of the vespertilionid bats in the yangochiropterans, rather than that of the rhinolophoids 632
[49,54]. Given this, laryngeal development may also be divergent within lineages with different 633
echolocation capabilities, as well as the petrosal and stylohyal (Nojiri et al., 2021b). The 634
molecular study detected positive selection of the hearing genes of the ancestral branches of bats, 635
suggesting that the common ancestor of bats employed primitive sonar, such as the lingual 636

echolocation of the genus Rousettus, and then rhinolophoids and yangochiropterans acquired 637
sophisticated laryngeal echolocation [82]. To date, there have been few studies on laryngeal 638
development in bats, especially in pteropodids and yangochiropterans, making this hypothesis 639
difficult to validate. This study reveals several key characteristics involved in the generation of 640
echolocation pulses. We identified significant differences in the morphology of the mineralized 641
laryngeal cartilage and the presence of the echolocation pulse between P0 and P14 specimens. 642
For example, the emission of the echolocation pulse was confirmed after P14, at which the lateral 643
tracheal chambers started those mineralization (Fig. 7C, D), possibly contributing to the 644
acquisition of the echolocation pulse production capability. Given this, the macroevolutionary 645
changes in the cricoid cartilage and acquisition of tracheal chambers might have evolved to some 646
extent with the acquisition of sophisticated biosonar in horseshoe bats. In subsequent studies, 647
investigations of the commonality and disparity of hyolaryngotracheal morphology in extinct and 648
extant bats would provide key insights into the evolutionary origins of laryngeal echolocation and 649
diversification of biosonar in bats. 650

Conclusions 651
Three-dimensional reconstruction of the hyolaryngeal apparatus revealed the unique morphology 652
of the cricoid cartilage, intrinsic laryngeal muscle, and cranial laryngeal nerve in horseshoe bats. 653
The present study suggests that the sagittal crest of the cricoid cartilage and the separated dorsal 654
cricoarytenoid muscles may be key features involved in echolocation pulse generation. 655
Laryngeally echolocating bats are characterized by hypertrophied intrinsic laryngeal muscles, 656
including the cricothyroid muscles, with superfast contractions. We provide a new perspective 657
that the hypertrophied cricothyroid muscle of horseshoe bats comprises homologous muscles in 658
mammals and possibly novel components acquired in rhinolophids and hipposiderids, which may 659
have served as key innovations in the adaptive radiation of these HDC bats. This drastic 660
morphological innovation is possibly in concert with the acquisition of tracheal chambers, 661
allowing them to conduct sophisticated biosonar with high accuracy for target localization. In 662
addition, we found that the maximum frequency of the echolocation pulse was significantly 663
correlated with mineralization of the cricoid and arytenoid cartilages. Given the acoustic diversity 664
in laryngeally echolocating bats, it is possible that the postnatal mineralization of each laryngeal 665
cartilage may independently regulate each acoustic property. Future studies integrating the 666
hyolaryngeal development of pteropodids, yangochiropterans, and other rhinolophoids will offer 667
insights into the relationship between variations in sound properties and hyolaryngeal 668
morphology and improve our understanding of how morphological changes in the hyolaryngeal 669
apparatus shape acoustic diversity in bats. 670
671

Methods 672
673
Data acquisition 674
A total of 32 specimens of the genus Rhinolophus species (R. pusillus, n = 28; R. malayanus, n = 675
4) and ten specimens of C57/BL6 mice (Mus musculus) were studied. The fetal and postnatal 676
specimens of R. pusillus were collected by the authors in Japan and Vietnam. Rhinolophus pusillus 677
from Japan was collected under the permission granted by the Gosen Local Government in Japan 678
(No. 1-4-2021) and is stored at the University of Tsukuba (UT). The fetal specimens of R. 679
malayanus were collected by the authors during field sampling in Vietnam under permission 680
granted by the An Giang Province People’s Committee in Vietnam (No. 497/VPUBND-NC and 681
No. 5893/VPUBND-KTN) and the Vietnam Administration of Forest, belonging to the Ministry 682
of Agriculture and Rural Development (No. 1072/TCLN-BTTN and No.326/TCLN-BTTN). 683
Rhinolophus pusillus from Japan and Vietnam are considered to compose a single species, and 684
geographic variation of the hyolaryngeal morphology was not detected between individuals from 685
Japan and those from Vietnam. Our postnatal samples of R. pusillus comprehensively covered the 686
overall ontogeny of the echolocation pulse, but some stages of prenatal specimens were lacking 687
to describe the ossification sequence of the hyolaryngeal bony components. Therefore, fetal 688
specimens of R. malayanus, both of which are members of the same genus Rhinolophus [83], 689
were additionally included. The examined specimens of R. pusillus and R. malayanus were first 690
confirmed to show no significant difference in qualitative hyolaryngeal morphology. Animal 691
ethics for mice sampling and experiments were approved at Tokyo Medical and Dental University, 692
Japan (A2019-060C3 and A2021-198A). All experiments were conducted following the ARRIVE 693
guidelines following the Japanese law on animal welfare. All pregnant individuals captured in 694
this study were euthanized by isoflurane inhalation. Fetal bat specimens were staged following 695
Cretekos Stage (CS) [84]. The postnatal stage of bats was estimated by the time when the first 696

birth was observed in the reproductive colony. The details of all specimens are summarized in 697
Table 1. Fetal specimens examined in this study include stages from CS16 to CS24 (R. pusillus: 698
CS16, 17, 18, 19, 20, 21, 22, 23, 24; R. malayanus: CS18, 20.5, 22). Postnatal specimens, all of 699
which were R. pusillus, covered postnatal day 0, 14, 21, 30, 45, and adult. One adult individual 700
of R. pusillus was loaned from the curatorial collection held in Kanagawa Prefectural Museum of 701
Natural History (KPMNH), and all samples of R. malayanus from the curatorial collection held 702
in the Institute of Ecology and Biological Resources, Vietnamese Academy of Science and 703
Technology (IEBR). All mouse fetuses from E14.5 to E18.5 and postnatal specimens from P0 to 704
adult (n = 10) were collected at Tokyo Medical and Dental University (TMDU). Pregnant mice 705
were sacrificed by cervical dislocation. The collected specimens were fixed using a mixture of 706
ethyl alcohol: acetic acid: formalin (6:3:1) and stored in 70% ethyl alcohol. Specimens used in 707
this study are summarized in Table 1. 708
709
Three-dimensional reconstruction of hyolaryngeal morphology 710
To investigate the development of the hyolaryngeal apparatus in the fetal and postnatal stages, we 711
used microcomputed tomography (micro-CT) at University of Tsukuba (inspeXio SMX-90CT 712
Plus, Shimadzu Corporation, Tokyo) and University Museum of University of Tokyo (TXS225-713
ACTIS, Tesco, Tokyo) with a 70 kV source voltage and 100 mA source current. To visualize the 714
cartilage and muscle for postnatal individuals (P0, 14, 21, 30, 45, and adult), the diffusible iodine-715
based contrast-enhanced CT technique [51] was conducted. This method has been shown to 716
effectively allow us to non-destructively investigate the morphology of bat soft tissues in detail 717
[36,37,85–88]. Specimens were stained with 1% iodine in ethanol for at least 48 hours and up to 718
14 days. The morphology of each hyolaryngeal component was manually segmented from the 719
aligned serial tissue sections (Fig. 12) and micro-CT images (Fig. 13) in Amira 5.3. 720
721

Histology 722
Serial tissue sections were prepared to reconstruct the detailed morphology that is still difficult to 723
visualize by the diffusible iodine-based contrast-enhanced CT. Three stages of bats (CS17, 724
CS20.5, CS22) and mice (E14.5, E16.5, E18.5) were dehydrated in ethanol series and embedded 725
in paraffin wax. Sections were cut at 6 μm for the CS17 bat and E14.5 mouse, and 7 μm for the 726
other stages. For immunohistochemistry (IHC) staining of CS22 of R. malayanus and E18.5 of 727
mice, mouse anti-acetylated tubulin (mouse monoclonal anti-acetylated tubulin, no. T7451; 728
Sigma-Aldrich Japan, Tokyo) was used to visualize the innervations. For CS17 of R. pusillus, 729
anti-Sox9 (rabbit polyclonal antibody, no. AB5535; Sigma-Aldrich Japan, Tokyo) was used to 730
visualize the chondrifications. The immunoreaction was visualized by the secondary antibody 731
conjugated with the anti-mouse IgG-Biotin antibody. The sections were then stained with Alcian 732
blue and hematoxylin to identify the cartilages and muscles following the standard staining 733
protocols. The obtained serial sections were aligned and then each targeted organ was manually 734
segmented using Amira 5.3 (Visage Imaging GmbH). 735
736
737
Measurements 738
To examine the allometric growth of the mineralized components against overall skull size in 739
horseshoe bats, the skull size and volume of the mineralized laryngeal cartilages (cricoid, 740
arytenoid, thyroid cartilages, lateral, and dorsal tracheal chambers) were measured using the 741
Surface Area Tool and 3D Length Tool in Amira 5.3. As the index of overall skull size [89], the 742
geometric mean (GM) was calculated from the skull length (SL), skull height (SH), and skull 743
width (SW) as: 744
745
𝐺𝑀3=𝑆𝐿 ∗𝑆𝐻 ∗𝑆𝑊 746

747
The cubed GM (GM3) was used as the index of the skull size in our allometric analyses. Skull 748
length (SL) was measured as the distance from the rostral tip of the premaxilla to the ventral end 749
of the interparietal in the sagittal plane. Skull height (SH) was measured as the distance from the 750
midpoint of the cranial tip of the left and right parietals to the rostral tip of the basioccipital. Skull 751
width (SW) was taken as the distance of the bilateral tip of the squamosal. 752
753
Bioacoustics 754
Echolocation pulses of P0 (n = 3), P14 (n = 3), P21 (n = 1), P30 (n = 2), P45 (n = 1), and adult (n 755
= 1) individuals were recorded, and sonograms were obtained during the field sampling using 756
Echo Meter Touch 2 Pro (Wildlife Acoustics Inc., Concord, MA). The mean of the sound 757
frequency of the first harmonic (CF1) and second harmonic (CF2) pulse of the constant frequency 758
(CF) was taken as the index for the capability of the biosonar pulse emission using Kaleidoscope 759
Pro Analysis Software (Wildlife Acoustics Inc., Concord, MA). 760
761
Ossification heterochrony 762
To examine the temporal specificity of the hyolaryngeal ossification in horseshoe bats, the relative 763
ossification timing of the hyolaryngeal components (arytenoid cartilage, basihyal, ceratohyal, 764
corniculate cartilage, cricoid cartilage, epihyal, stylohyal, thyrohyal, thyroid cartilage, dorsal and 765
lateral tracheal chambers) was estimated, The ossification sequence of 31 craniocervical elements 766
(alisphenoid, arytenoid cartilage, basihyal, basioccipital, basisphenoid, ceratohyal, corniculate 767
cartilage, cricoid cartilage, dentary, dorsal and lateral tracheal chambers, ectotympanic, epihyal, 768
exoccipital, frontal, goniale, jugal, lacrimal, maxilla, nasal, orbitosphenoid, palatine, parietal, 769
petrosal, premaxilla, presphenoid, pterygoid, squamosal, stylohyal, supraoccipital, thyroid 770
cartilage, thyrohyal, and vomer) was documented for Rhinolophus species and M. musculus. To 771

examine the relative ossification timing of each bony component and allow interspecific 772
comparison, the rank of each ossification event was scaled as: 773
774
(𝑟 − 1)
(𝑟
𝑚𝑎𝑥 − 1) 775
776
in which r is the absolute rank of each ossification event, and rmax is the total number of ranks for 777
each species. The scaled relative values are distributed between 0 and 1. As the sequence 778
resolution can affect the results in this approach, >3 ranks were documented for each species 779
following [90]. 780
781
Statistics 782
All statistical analyses were performed in PAST4. To examine the size allometry of the volume 783
of the mineralized laryngeal cartilages, the reduced major axis (RMA) regression analysis was 784
conducted. To rescale the data and handle them on a common scale, the volume of the cricoid 785
(CV), thyroid (TV), arytenoid (AV), and lateral tracheal chamber (CH) were log10-transformed 786
and regressed against log10-transformed GM3. Furthermore, the regression line of the maximum 787
frequency of the log10-transformed CF1 and CF2 components against log10-transformed GM3 788
was also obtained. Given that the RMA provides the best-estimated regression between every 789
population from which the sample is selected when the error variance is unknown and is not 790
affected by the correlation coefficient of samples, the RMA analysis is considered to be more 791
appropriate than other regression models [91,92]. 792
To investigate the relationship between the pulse frequency and mineralization of the 793
laryngeal cartilages, the generalized linear model (GLM) with identity link analyses was 794
conducted. GLM analysis allows us to accommodate different types of response variables using 795

the link function for various probability distributions and include the multiple independent 796
variables in model, thus controlling for potential confounding factors that may influence the 797
relationship between pulse peak frequency and mineralized volume of the laryngeal cartilages 798
[93]. In the GLM analysis, the pulse frequency of CF1 and CF2 was set as the response variable, 799
and the log-transformed volume of the mineralized cricoid cartilage, thyroid cartilage, arytenoid 800
cartilage, lateral tracheal chamber, and dorsal tracheal chamber was as the explanatory variable. 801
Since the pulse peak frequency is a continuous variable, it was assumed to follow a normal 802
distribution. The volume of each mineralized laryngeal cartilage was divided by the estimated 803
skull volume to remove the effect of the skull size growth (relative cricoid cartilage volume: RCV, 804
relative thyroid cartilage volume: RTV, relative arytenoid cartilage volume: RAV, and relative 805
lateral tracheal chamber volume: RCHV). 806
807
Anatomical abbreviations 808
ac arytenoid cartilage 809
ba basihyal 810
cc cricoid cartilage 811
ce ceratohyal 812
cln cranial laryngeal nerve 813
co corniculate cartilage 814
cp corniculate process of the arytenoid cartilage 815
ct cricothyroid muscle 816
dca dorsal cricoarytenoid muscle 817
dch dorsal tracheal chamber 818
eg epiglottis 819
ep epihyal 820

g glottis 821
hc hyoid complex 822
lca lateral cricoarytenoid muscle 823
lch lateral tracheal chamber 824
mw muscular wing of the cricoid cartilage 825
pt protuberance of the thyroid cartilage 826
rln recurrent laryngeal nerve 827
sc sagittal crest of the cricoid cartilage 828
st stylohyal 829
t trachea 830
ta thyroarytenoid muscle 831
tc thyroid cartilage 832
tcac thyroid caudal cornua 833
tcrc thyroid cranial cornua 834
th thyrohyal 835
thm thyrohyoid muscle 836
tr tracheal ring 837
vn vagus nerve 838
vp vocal process of the arytenoid cartilage 839
840
841

Declarations 842
Not applicable. 843
844
Ethics approval and consent to participate 845
Animal ethics for the experiments were approved by the Tokyo Medical and Dental University, 846
Japan (A2019-060C3 and A2021-198A). 847
848
Consent for publication 849
Not applicable. 850
851
Availability of data and materials 852
All data supporting the findings of this study are available from the corresponding authors. 853
854
Competing interests 855
The authors declare no conflicts of interest. 856
857
Funding 858
JSPS (19J20608, 22J00127, and 23K14255) to N. T.; JSPS 22K06786 to M. T.; JSPS 22K06337 859
to T. F.; JSPS (18KK0207, 18K19359, 21H02546, 21K19291, and 22KK0101), JST 860
(JPMJFR2148), the Naito Foundation, and JRPs-LEAD with DFG (JPJSJRP20181608) to D.K. 861
862
Author contributions 863
Taro Nojiri, Masaki Takechi and Daisuke Koyabu conceived the study. Taro Nojiri, Dai Fukui, 864
Vuong Tan Tu, Toshiko Furutera, Fumiya Meguro, and Daisuke Koyabu collected rhinolophid 865
samples from the fields. Taro Nojiri and Daisuke Koyabu collected pulse data. Takechi and 866

Furutera collected mouse embryos. Taro Nojiri, Masaki Takechi, Toshiko Furutera, and Daisuke 867
Koyabu scanned the specimens using micro-CT and stained the serial tissue sections. Taro Nojiri 868
and Daisuke Koyabu drafted the manuscript. Taro Nojiri conducted all the analyses. All authors 869
provided insights and critically revised the manuscript. 870
871
Acknowledgements 872
We are grateful to Ayumi Shimanuki for the technical assistance. We thank Tatsuya Hirasawa, 873
Daichi Suzuki, and Yoshitaka Tanaka for their insightful comments. We thank the Gosen Local 874
Government, An Giang Province People’s Committee of Vietnam, and the Vietnam 875
Administration of Forest, which belongs to the Ministry of Agriculture and Rural Development, 876
for their research permits and support. We thank Satoshi Suzuki of the KPMNH for allowing 877
access to their collections. 878
879
Authors’ information 880
ORCID 881
Taro Nojiri: https://orcid.org/0000-0002-9984-3467
882
Masaki Takechi: http://orcid.org/0000-0001-7279-9243
883
Toshiko Furutera: http://orcid.org/0000-0002-3617-2360
884
Nicolas L.M. Brualla: http://orcid.org/0000-0003-1367-0778
885
Dai Fukui: http://orcid.org/0000-0002-5449-4283
886
Vuong Tan Tu: http://orcid.org/0000-0002-5915-865X
887
Daisuke Koyabu: http://orcid.org/0000-0002-4087-7742 888
889

LEGENDS 890
891
Table 1. All specimens examined in this study. 892
893
species
ID
storage
stage
bone
micro-CT
dice
CT
tissue
section
Mus musculus
NT22-002
TMDU
E14.5
x
Mus musculus
NT22-003
TMDU
E16.5
x
Mus musculus
JP19-001
TMDU
E17.5
x
Mus musculus
NT22-004
TMDU
E18.0
x
Mus musculus
NT22-005
TMDU
postnatal day 0
x
Mus musculus
NT22-011
TMDU
postnatal day 0
x
Mus musculus
NT22-007
TMDU
postnatal day 7
x
Mus musculus
NT22-008
TMDU
postnatal day 14
x
Mus musculus
NT22-012
TMDU
postnatal day 14
x
Mus musculus
NT22-010
TMDU
adult
x
Rhinolophus
pusillus
JP20-052
UT
CS16
x
Rhinolophus
pusillus
JP20-044
UT
CS17
x
Rhinolophus
pusillus
JP20-055
UT
CS17
x
Rhinolophus
pusillus
JP20-039
UT
CS18
x
Rhinolophus
pusillus
QN043
IEBR
CS19
x
Rhinolophus
pusillus
QN041
IEBR
CS20
x
Rhinolophus
pusillus
B160413-
4
IEBR
CS21
x
Rhinolophus
pusillus
JP21-032
UT
CS22
x
x
Rhinolophus
pusillus
XL2016-
25
IEBR
CS22
x
Rhinolophus
pusillus
VN17-
299
IEBR
CS23
x
Rhinolophus
pusillus
B200413-
7
IEBR
CS24
x
Rhinolophus
pusillus
JP21-047
UT
postnatal day 0
x
Rhinolophus
pusillus
JP21-073
UT
postnatal day 14
x

Rhinolophus
pusillus
JP21-074
UT
postnatal day 14
x
Rhinolophus
pusillus
JP21-075
UT
postnatal day 14
x
x
Rhinolophus
pusillus
JP22-067
UT
postnatal day 14
x
Rhinolophus
pusillus
JP22-068
UT
postnatal day 14
x
Rhinolophus
pusillus
JP22-082
UT
postnatal day 21
x
Rhinolophus
pusillus
JP21-083
UT
postnatal day 21
x
Rhinolophus
pusillus
JP21-091
UT
postnatal day 21
x
Rhinolophus
pusillus
JP21-100
UT
postnatal day 30
x
x
Rhinolophus
pusillus
JP21-101
UT
postnatal day 30
x
Rhinolophus
pusillus
JP21-104
UT
postnatal day 45
x
x
Rhinolophus
pusillus
JP21-025
UT
adult
x
Rhinolophus
pusillus
KPMNH2
937
IEBR
adult
x
Rhinolophus
malayanus
VN20-
032
IEBR
CS18
x
Rhinolophus
malayanus
VN20-
054
IEBR
CS18
x
Rhinolophus
malayanus
VN19-
057
IEBR
CS20.5
x
Rhinolophus
malayanus
VN19-
029
IEBR
CS22
x
894

Table 2. Summary of measurements in each specimen in this study. 895
896
897
species ID stage
GM3 (mm3)first harmonic (kHz)
second harmonic
(kHz)
cricoid cartilage
volume (mm3)
thyroid cartilage
volume (mm3)
arytenoid cartilage
volume (mm3)
lateral tracheal
chamber volume
(mm3)
Rhinolophus cornutus
JP21-045 postnatal day 0 572.26 0.80 0.18 0.11
Rhinolophus cornutus
JP21-046 postnatal day 0 520.08 0.62 0.20 0.11
Rhinolophus cornutus
JP21-047 postnatal day 0 586.54 0.86 0.27 0.17
Rhinolophus cornutus
JP21-073 postnatal day 14 623.39 49.99 99.98 0.85 0.31 0.19 0.21
Rhinolophus cornutus
JP21-074 postnatal day 14 645.35 49.58 99.20 0.93 0.37 0.22 0.27
Rhinolophus cornutus
JP21-075 postnatal day 14 595.82 49.90 99.93 0.86 0.34 0.23 0.19
Rhinolophus cornutus
JP21-083 postnatal day 21 644.23 50.67 101.23 0.87 0.35 0.23 0.32
Rhinolophus cornutus
JP21-097 postnatal day 30 681.38 51.28 102.67 0.99 0.35 0.29 0.32
Rhinolophus cornutus
JP21-100 postnatal day 30 681.38 50.89 101.83 1.01 0.41 0.27 0.31
Rhinolophus cornutus
JP21-101 postnatal day 30 671.22 50.92 101.92 0.95 0.36 0.24 0.32
Rhinolophus cornutus
JP21-104 postnatal day 45 701.12 51.69 103.37 1.06 0.38 0.31 0.36
Rhinolophus cornutus
JP21-103 adult 54.14 108.25
Rhinolophus cornutus
NHKM2937- a dult 704.82 1.37 0.53 0.34 0.57

Table 3. Basic statistics of the allometric analyses of log10-transformed volume of the mineralized laryngeal cartilages or echolocation pulse, against 898
log10-transformed geometric mean (GM). 899
900
Slope (95%CI)
Intercept (95% CI)
r
p-value
cricoid cartilage
2.01 (1.21 ~ 2.83)
-5.68 (-7.95 ~ -3.43)
0.90
p < 0.0001
thyroid cartilage
3.20 (1.82 ~ 4.39)
-9.46 (-12.82 ~ -5.59)
0.88
p < 0.001
arytenoid cartilage
3.93 (2.71 ~ 4.88)
-11.68 (-14.36 ~ -8.23)
0.93
p < 0.001
tracheal chamber
5.66 (3.14 ~ 7.69)
-16.47 (-22.21 ~ -9.36)
0.88
p < 0.01
CF1
0.47 (0.04 ~ 0.73)
0.39 (-0.36 ~ 1.61)
0.76
0.0165
CF2
0.46 (0.05 ~ 0.73)
0.70 (-0.04 ~ 1.88)
0.76
0.0166
901

Table 4. Basic statistics of the generalized linear model analyses between log10-transformed volume of the mineralized laryngeal cartilages and log10-902
transformed peak frequency of CF1 or CF2. 903
904
CF1
laryngeal cartilage
slope
standard error of slope
intercept
standard error of intercept
p-value
cricoid cartilage
0.21
0.04
2.31
0.12
p < 0.0001
thyroid cartilage
0.15
0.06
2.21
0.19
p < 0.01
arytenoid cartilage
0.15
0.04
2.22
0.14
p < 0.001
lateral tracheal chamber
0.09
0.02
1.99
0.06
p < 0.0001
CF2
laryngeal cartilage
slope
standard error
intercept
standard error of intercept
p-value
cricoid cartilage
0.21
0.04
2.61
0.12
p < 0.0001
thyroid cartilage
0.15
0.06
2.51
0.18
p < 0.01
arytenoid cartilage
0.15
0.04
2.53
0.13
p < 0.0001
lateral tracheal chamber
0.08
0.02
2.29
0.06
p < 0.0001
905

906
Fig. 1. The gross anatomy of the hyolaryngeal apparatus in bats (postnatal day 0 of Rhinolophus 907
pusillus) and mice (postnatal day 14 of Mus musculus) reconstructed with the diffusible iodine-908
based contrast-enhanced CT scanning. Left side of the hyolaryngeal morphology is shown in 909
lateral view. Scale bars = 1 mm. See text for abbreviations. 910

911
Fig. 2. The whole morphology of the intrinsic laryngeal muscle and innervations of the cranial 912
laryngeal nerve and recurrent laryngeal nerve of CS22 stage in Rhinolophus malayanus and E18.5 913
stage in Mus musculus. A: Dorsal view of the hyolaryngeal cartilages, intrinsic laryngeal muscles, 914
and vagus nerve of CS22 fetus of R. malayanus. B: Lateral view of the hyolaryngeal cartilages, 915
intrinsic laryngeal muscles, and vagus nerve of CS 22 fetus of R. malayanus. C: Dorsal view of 916
the intrinsic laryngeal muscles and vagus nerve of CS22 fetus of R. malayanus. D: Lateral view 917
of the intrinsic laryngeal muscles and vagus nerve of CS22 fetus of R. malayanus. E: Dorsal view 918
of the hyolaryngeal cartilages, intrinsic laryngeal muscles, cranial laryngeal nerve, and recurrent 919
laryngeal nerve of E18.5 fetus of M. musculus. F: Lateral view of the hyolaryngeal cartilages, 920
intrinsic laryngeal muscles, cranial laryngeal nerve, and recurrent laryngeal nerve of E18.5 fetus 921
of M. musculus. G: Dorsal view of the intrinsic laryngeal muscles, cranial laryngeal nerve, and 922
recurrent laryngeal nerve of E18.5 fetus of M. musculus. H: Lateral view of the intrinsic laryngeal 923
muscles, cranial laryngeal nerve, and recurrent laryngeal nerve of E18.5 fetus of M. musculus. 924
The cranial laryngeal nerve and recurrent laryngeal nerve were visualized using the 925
immunohistochemistry of the acetylated tubulin antibody. The left side of the hyolaryngeal 926
morphology is shown in lateral view. See text for abbreviations. 927
928

929
Fig. 3. Prenatal development of the hyoid apparatus in horseshoe bats (CS17, CS20.5, CS22) and 930
Mus musculus (E14.5, E16.5, E18.5), three-dimensionally reconstructed using serial tissue 931
sections. A: Rostral view of the hyolaryngeal apparatus of CS17 fetus of the horseshoe bat. B: 932
Lateral view of the hyoid components of CS17 fetus of Rhinolophus pusillus. C: Rostral view of 933
the hyolaryngeal apparatus of CS20.5 fetus of R. malayanus. D: Lateral view of the hyoid 934
components of CS20.5 fetus of R. malayanus. E: Rostral view of the hyolaryngeal apparatus of 935
CS22 fetus of R. malayanus. F: Lateral view of the hyoid components of CS22 fetus of R. 936
malayanus. G: Rostral view of the hyolaryngeal apparatus of E14.5 fetus of M. musculus. H: 937
Lateral view of the hyoid components of E14.5 fetus of M. musculus. I: Rostral view of the 938
hyolaryngeal apparatus of E16.5 fetus of M. musculus. J: Lateral view of the hyoid components 939
of E16.5 fetus of M. musculus. K: Rostral view of the hyolaryngeal apparatus of E18.5 fetus of 940
M. musculus. L: Lateral view of the hyoid components of E18.5 fetus of M. musculus. The tracheal 941
chambers and tracheal rings of the horseshoe bats start those chondrifications at CS17 and CS20.5, 942
respectively. The left side of the hyolaryngeal morphology is shown in lateral view. See text for 943
abbreviations. 944
945

946
Fig. 4. Prenatal laryngeal development of horseshoe bats and laboratory mice, three-947
dimensionally reconstructed from serial tissue sections. See text for abbreviations. A: Ventral 948
view of the hyolaryngeal apparatus of CS17 fetus of Rhinolophus pusillus. B: Dorsal view of the 949
hyolaryngeal apparatus of CS17 fetus of R. pusillus. C: Caudal view of the hyolaryngeal apparatus 950
of CS17 fetus of R. pusillus. D: Ventral view of the hyolaryngeal apparatus of CS20.5 fetus of R. 951
pusillus. E: Dorsal view of the hyolaryngeal apparatus of CS20.5 fetus of R. malayanus. F: Caudal 952
view of the hyolaryngeal apparatus of CS20.5 fetus of R. malayanus. G: Ventral view of the 953
hyolaryngeal apparatus of CS22 fetus of R. pusillus. H: Dorsal view of the hyolaryngeal apparatus 954
of CS22 fetus of R. malayanus. I: Caudal view of the hyolaryngeal apparatus of CS22 fetus of R. 955
malayanus. J: Ventral view of the hyolaryngeal apparatus of E14.5 fetus of Mus musculus. K: 956
Dorsal view of the hyolaryngeal apparatus of E14.5 fetus of M. musculus. L: Caudal view of the 957
hyolaryngeal apparatus of E14.5 fetus of M. musculus. M: Ventral view of the hyolaryngeal 958
apparatus of E16.5 fetus of M. musculus. N: Dorsal view of the hyolaryngeal apparatus of E16.5 959
fetus of M. musculus. O: Caudal view of the hyolaryngeal apparatus of E16.5 fetus of M. musculus. 960
P: Ventral view of the hyolaryngeal apparatus of E18.5 fetus of M. musculus. Q: Dorsal view of 961
the hyolaryngeal apparatus of E18.5 fetus of M. musculus. R: Caudal view of the hyolaryngeal 962
apparatus of E18.5 fetus of M. musculus. 963
964

965
Fig. 5. The three-dimensional reconstruction of the fetal tracheal development and histological 966
observations in horseshoe bats and laboratory mice using immunohistochemistry. A-1: the 967
tracheal anatomy of CS17 fetal specimen of Rhinolophus pusillus. A-2: the histological section 968
of the hyolarynx of CS17 fetal specimen of R. pusillus, stained by hematoxylin and 969
immunohistochemistry of Sox9. B-1: the tracheal anatomy of CS22 fetal specimen of R. 970
malayanus. B-2 to B-4: the histological sections of the hyolarynx of CS22 fetal specimen of R. 971
malayanus, stained by alcian blue, hematoxylin, and immunohistochemistry of acetylated tubulin 972
antibody. C-1: the tracheal anatomy of E14.5 fetal specimen of Mus musculus. C-2: the 973
histological section of the hyolarynx of E14.5 fetal specimen of M. musculus, stained by alcian 974
blue, hematoxylin, and immunohistochemistry of acetylated tubulin antibody. D-1: the tracheal 975
anatomy of E16.5 fetal specimen of M. musculus. D-2: the histological section of the hyolarynx 976
of E16.5 fetal specimen of M. musculus, stained by alcian blue, hematoxylin, and 977
immunohistochemistry of acetylated tubulin antibody. E-1: the tracheal anatomy of E18.5 fetal 978
specimen of M. musculus. E-2: the histological section of the hyolarynx of E18.5 fetal specimen 979
of M. musculus, stained by alcian blue, hematoxylin, and immunohistochemistry of acetylated 980
tubulin antibody. Scale bars = 500 μm. See text for abbreviations. 981
982

983
Fig. 6. Three-dimensional reconstruction of the pre- and postnatal ossification and mineralization 984
of the hyolaryngeal apparatus in horseshoe bats. Scale bars = 1 mm. See text for abbreviations. 985
CS21: fetal specimen of Rhinolophus pusillus. The stylohyal is ossified. CS22: fetal specimen of 986
R. pusillus. The thyrohyal and epihyal are ossified. CS24: fetal specimen of R. pusillus. The 987
basihyal and thyrohyal are ossified. The arytenoid, cricoid, and thyroid cartilages are mineralized. 988
P0: postnatal day 0 specimen of R. pusillus. The sagittal crest of the cricoid cartilage is 989
mineralized. P14: postnatal day 14 specimen of R. pusillus. The lateral tracheal chambers are 990
mineralized. P21: postnatal day 21 specimen of R. pusillus. The dorsal tracheal chambers are 991
mineralized. Adult: adult specimen of R. pusillus. 992

993
Fig. 7. Three-dimensional reconstruction of postnatal ossification and mineralization of the 994
hyolaryngeal apparatus in horseshoe bats (Rhinolophus pusillus) and laboratory mice (Mus 995
musculus) based on micro-CT scanning. Scale bars = 1 mm. See text for abbreviations. A: Lateral 996
view of the hyolaryngeal apparatus of the postnatal day 0 specimen of R. pusillus. B: Dorsal view 997
of the hyolaryngeal apparatus of the postnatal day 0 specimen of R. pusillus. C: Lateral view of 998
the hyolaryngeal apparatus of the postnatal day 14 specimen of R. pusillus. D: Dorsal view of the 999
hyolaryngeal apparatus of the postnatal day 14 specimen of R. pusillus. E: Lateral view of the 1000
hyolaryngeal apparatus of the adult specimen of R. pusillus. F: Dorsal view of the hyolaryngeal 1001
apparatus of the adult specimen of R. pusillus. G: Lateral view of the hyolaryngeal apparatus of 1002
the postnatal day 0 specimen of M. musculus. H: Dorsal view of the hyolaryngeal apparatus of 1003
the postnatal day 0 specimen of M. musculus. I: Lateral view of the hyolaryngeal apparatus of the 1004
postnatal day 14 specimen of M. musculus. J: Dorsal view of the hyolaryngeal apparatus of the 1005
postnatal day 14 specimen of M. musculus. K: Lateral view of the hyolaryngeal apparatus of the 1006
adult specimen of M. musculus. L: Dorsal view of the hyolaryngeal apparatus of the adult 1007
specimen of M. musculus. 1008
1009

1010
Fig. 8. The relative ossification and mineralization timing of the bony and cartilaginous 1011
components of the hyolaryngeal apparatus in horseshoe bats (red) and laboratory mice (white). 1012
See text for abbreviations. See text for abbreviations. 1013
1014

1015
Fig. 9. The regression analyses based on the reduced major axis between the mineralized volume 1016
of the laryngeal cartilages or pulse peak frequency against skull size (cubed geometric mean). A: 1017
log10-transformed volume of the mineralized cricoid cartilage and log10-transformed GM3; B: 1018
log10-transformed volume of the mineralized thyroid cartilage and log10-transformed GM3; C: 1019
log10-transformed volume of the mineralized arytenoid cartilage and log10-transformed GM3; D: 1020
log10-transformed volume of the mineralized tracheal chamber and log10-transformed GM3; E: 1021
log10-transformed CF2 peak frequency and log10-transformed GM; F: log10-transformed CF1 peak 1022
frequency and log10-transformed GM3. CV: cricoid cartilage volume; TV: thyroid cartilage 1023
volume; AV: arytenoid cartilage volume; CHV: tracheal chambers volume; GM3: cubed geometric 1024
mean of the skull height (SH), skull length (SL), and skull width (SW); . Note that plots for P0 1025
individuals are not shown in D, E, and F, as mineralization and echolocation pulse are absent at 1026
this stage. 1027

1028
Fig. 10. Ontogeny of the hyolaryngeal apparatus morphology and the echolocation pulse in 1029
Rhinolophus pusillus. 1030
1031
1032

1033
Fig. 11. GLM analyses between the pulse peak frequency and relative volume of the mineralized 1034
laryngeal cartilages. Each volume was normalized by the cubed geometric mean of the skull 1035
height (SH), skull length (SL), and skull width (SW). 1036

1037
Fig. 12. The three-dimensional reconstruction procedure of the hard tissue and soft tissue of the 1038
hyolaryngeal apparatus in postnatal day 30 of Rhinolophus pusillus. A: One example of the 1039
tomographic image obtained from micro-CT scanning without staining. B: Three-dimensionally 1040
reconstruction process of the morphology of the hyolaryngeal cartilages with the tomographic 1041
images. C: Three-dimensionally reconstructed morphology of the hyolaryngeal cartilages. D: One 1042
example of the tomographic image obtained from diffusible iodine-based contrast-enhanced CT 1043
method. E: Three-dimensionally reconstruction process of the morphology of the hyolaryngeal 1044
cartilages with the tomographic images. F: Three-dimensionally reconstructed morphology of the 1045
hyolaryngeal cartilages, intrinsic laryngeal muscles, and vagus nerve (cranial nerve X). Left side 1046
of the hyolaryngeal morphology is shown. 1047

1048
Fig. 13. The three-dimensional reconstruction procedure of the gross anatomy of the hyolarynx 1049
of the bat fetus (CS22; Rhinolophus pusillus) with serial tissue sections. All sections were 1050
immunostained with the acetylated tubulin antibody to visualize the nerves. Left side of the 1051
hyolarynx is shown in the lateral view. After immunostaining, all sections were additionally 1052
stained with alcian blue and hematoxylin. Scale bars = 500 μm. See text for abbreviations. 1053

REFERENCES 1054
1055
1. Fenton MB. Evolution of Echolocation. In: Adams RA, Pedersen SC, editors. Bat Evolution, 1056
Ecology, and Conservation [Internet]. New York, NY: Springer New York; 2013. p. 47–70. 1057
Available from: https://doi.org/10.1007/978-1-4614-7397-8_3 1058
2. Ketten DR. The Marine Mammal Ear: Specializations for Aquatic Audition and Echolocation. 1059
In: Webster DB, Popper AN, Fay RR, editors. The Evolutionary Biology of Hearing [Internet]. 1060
New York, NY: Springer New York; 1992. p. 717–50. Available from: 1061
https://doi.org/10.1007/978-1-4612-2784-7_44 1062
3. O’Connell-Rodwell CE, Arnason BT, Hart LA. Seismic properties of Asian elephant (Elephas 1063
maximus) vocalizations and locomotion. J Acoust Soc Am [Internet]. 2000;108:3066–72. 1064
Available from: http://dx.doi.org/10.1121/1.1323460 1065
4. Charlton BD, Frey R, McKinnon AJ, Fritsch G, Fitch WT, Reby D. Koalas use a novel vocal 1066
organ to produce unusually low-pitched mating calls. Curr Biol [Internet]. 2013;23:R1035-6. 1067
Available from: http://dx.doi.org/10.1016/j.cub.2013.10.069 1068
5. Frey R, Riede T. The anatomy of vocal divergence in North American Elk and European red 1069
deer. J Morphol [Internet]. 2013;274:307–19. Available from: 1070
http://dx.doi.org/10.1002/jmor.20092 1071
6. Fitch WT. Vocal tract length and formant frequency dispersion correlate with body size in 1072
rhesus macaques. J Acoust Soc Am [Internet]. 1997;102:1213–22. Available from: 1073
http://dx.doi.org/10.1121/1.421048 1074
7. Nishimura T. Comparative morphology of the hyo-laryngeal complex in anthropoids: two steps 1075
in the evolution of the descent of the larynx. Primates [Internet]. 2003;44:41–9. Available from: 1076
http://dx.doi.org/10.1007/s10329-002-0005-9 1077
8. Nishimura T, Oishi T, Suzuki J, Matsuda K, Takahashi T. Development of the supralaryngeal 1078
vocal tract in Japanese macaques: implications for the evolution of the descent of the larynx. Am 1079
J Phys Anthropol [Internet]. 2008;135:182–94. Available from: 1080
http://dx.doi.org/10.1002/ajpa.20719 1081
9. Zhang YS, Takahashi DY, Liao DA, Ghazanfar AA, Elemans CPH. Vocal state change through 1082
laryngeal development. Nat Commun [Internet]. 2019;10:4592. Available from: 1083
http://dx.doi.org/10.1038/s41467-019-12588-6 1084

10. Bowling DL, Dunn JC, Smaers JB, Garcia M, Sato A, Hantke G, et al. Rapid evolution of the 1085
primate larynx? PLoS Biol [Internet]. 2020;18:e3000764. Available from: 1086
http://dx.doi.org/10.1371/journal.pbio.3000764 1087
11. Fitch WT, de Boer B, Mathur N, Ghazanfar AA. Monkey vocal tracts are speech-ready. Sci 1088
Adv [Internet]. 2016;2:e1600723. Available from: http://dx.doi.org/10.1126/sciadv.1600723 1089
12. Nishimura T, Tokuda IT, Miyachi S, Dunn JC, Herbst CT, Ishimura K, et al. Evolutionary loss 1090
of complexity in human vocal anatomy as an adaptation for speech. Science [Internet]. 1091
2022;377:760–3. Available from: http://dx.doi.org/10.1126/science.abm1574 1092
13. Landrau-Giovannetti N, Mignucci-Giannoni AA, Reidenberg JS. Acoustical and anatomical 1093
determination of sound production and transmission in West Indian (Trichechus manatus) and 1094
Amazonian (T. inunguis) manatees. Anat Rec [Internet]. 2014;297:1896–907. Available from: 1095
https://anatomypubs.onlinelibrary.wiley.com/doi/abs/10.1002/ar.22993 1096
14. Borgard HL, Baab K, Pasch B, Riede T. The shape of sound: a geometric morphometrics 1097
approach to laryngeal functional morphology. J Mamm Evol [Internet]. 2020;27:577–90. 1098
Available from: https://doi.org/10.1007/s10914-019-09466-9 1099
15. Riede T, Coyne M, Tafoya B, Baab KL. Postnatal development of the mouse larynx (Riede et 1100
al., 2020) [Internet]. 2020. Available from: 1101
https://asha.figshare.com/articles/journal_contribution/Postnatal_development_of_the_mouse_l1102
arynx_Riede_et_al_2020_/12735917 1103
16. He K, Liu Q, Xu D-M, Qi F-Y, Bai J, He S-W, et al. Echolocation in soft-furred tree mice. 1104
Science [Internet]. 2021;372. Available from: http://dx.doi.org/10.1126/science.aay1513 1105
17. Lesch R, Schwaha T, Orozco A, Shilling M, Brunelli S, Hofer M, et al. Selection on vocal 1106
output affects laryngeal morphology in rats. J Anat [Internet]. 2021;238:1179–90. Available from: 1107
http://dx.doi.org/10.1111/joa.13366 1108
18. Smith SK, Burkhard TT, Phelps SM. A comparative characterization of laryngeal anatomy in 1109
the singing mouse. J Anat [Internet]. 2021;238:308–20. Available from: 1110
http://dx.doi.org/10.1111/joa.13315 1111
19. Elemans CPH, Mead AF, Jakobsen L, Ratcliffe JM. Superfast muscles set maximum call rate 1112
in echolocating bats. Science [Internet]. 2011;333:1885–8. Available from: 1113
http://dx.doi.org/10.1126/science.1207309 1114

20. Håkansson J, Mikkelsen C, Jakobsen L, Elemans CPH. Bats expand their vocal range by 1115
recruiting different laryngeal structures for echolocation and social communication. PLoS Biol 1116
[Internet]. 2022;20:e3001881. Available from: http://dx.doi.org/10.1371/journal.pbio.3001881 1117
21. Teeling EC, Springer MS, Madsen O, Bates P, O’brien SJ, Murphy WJ. A molecular 1118
phylogeny for bats illuminates biogeography and the fossil record. Science [Internet]. 1119
2005;307:580–4. Available from: http://dx.doi.org/10.1126/science.1105113 1120
22. Jones G, Teeling EC. The evolution of echolocation in bats. Trends Ecol Evol [Internet]. 1121
2006;21:149–56. Available from: http://dx.doi.org/10.1016/j.tree.2006.01.001 1122
23. Yovel Y, Geva-Sagiv M, Ulanovsky N. Click-based echolocation in bats: not so primitive after 1123
all. J Comp Physiol A Neuroethol Sens Neural Behav Physiol [Internet]. 2011;197:515–30. 1124
Available from: http://dx.doi.org/10.1007/s00359-011-0639-4 1125
24. Boonman A, Bumrungsri S, Yovel Y. Nonecholocating fruit bats produce biosonar clicks with 1126
their wings. Curr Biol [Internet]. 2014;24:2962–7. Available from: 1127
http://dx.doi.org/10.1016/j.cub.2014.10.077 1128
25. Pedersen SC. Morphometric analysis of the Chiropteran skull with regard to mode of 1129
echolocation. J Mammal [Internet]. 1998;79:91–103. Available from: 1130
https://www.researchgate.net/publication/271791226 1131
26. Denny SP. The bat larynx. Scientific foundations of otolaryngology; 1976. 1132
27. Roberts LH. Variable resonance in constant frequency bats. J Zool [Internet]. 1972;166:337–1133
48. Available from: https://onlinelibrary.wiley.com/doi/10.1111/j.1469-7998.1972.tb03103.x 1134
28. Griffiths TA. Modification of m. cricothyroideus and the larynx in the Mormoopidae, with 1135
reference to amplification of high-frequency pulses. J Mammal. 1978;59:724–30. 1136
29. Hiryu S, Katsura K, Lin L-K, Riquimaroux H, Watanabe Y. Doppler-shift compensation in the 1137
Taiwanese leaf-nosed bat (Hipposideros terasensis) recorded with a telemetry microphone system 1138
during flight. J Acoust Soc Am [Internet]. 2005;118:3927–33. Available from: 1139
http://dx.doi.org/10.1121/1.2130940 1140
30. Habersetzer J, Storch G. Cochlea size in extant chiroptera and middle eocene 1141
microchiropterans from messel [Internet]. Naturwissenschaften. 1992. p. 462–6. Available from: 1142
http://dx.doi.org/10.1007/bf01139198 1143
31. Simmons NB, Seymour KL, Habersetzer J, Gunnell GF. Primitive Early Eocene bat from 1144

Wyoming and the evolution of flight and echolocation. Nature [Internet]. 2008;451:818–21. 1145
Available from: http://dx.doi.org/10.1038/nature06549 1146
32. Veselka N, McErlain DD, Holdsworth DW, Eger JL, Chhem RK, Mason MJ, et al. A bony 1147
connection signals laryngeal echolocation in bats. Nature [Internet]. 2010;463:939–42. Available 1148
from: http://dx.doi.org/10.1038/nature08737 1149
33. Davies KT, Maryanto I, Rossiter SJ. Evolutionary origins of ultrasonic hearing and laryngeal 1150
echolocation in bats inferred from morphological analyses of the inner ear. Front Zool [Internet]. 1151
2013;10:2. Available from: http://dx.doi.org/10.1186/1742-9994-10-2 1152
34. Carter RT, Adams RA. Postnatal ontogeny of the cochlea and flight ability in Jamaican fruit 1153
bats (Phyllostomidae) with implications for the evolution of echolocation. J Anat [Internet]. 1154
2015;226:301–8. Available from: http://dx.doi.org/10.1111/joa.12284 1155
35. Nojiri T, Werneburg I, Son NT, Tu VT, Sasaki T, Maekawa Y, et al. Prenatal cranial bone 1156
development of Thomas’s horseshoe bat (Rhinolophus thomasi): with special reference to petrosal 1157
morphology. J Morphol [Internet]. 2018;279:809–27. Available from: 1158
http://dx.doi.org/10.1002/jmor.20813 1159
36. Nojiri T, Fukui D, Werneburg I, Saitoh T, Endo H, Koyabu D. Embryonic staging of bats with 1160
special reference to Vespertilio sinensis and its cochlear development. Dev Dyn [Internet]. 1161
2021;250:1140–59. Available from: http://dx.doi.org/10.1002/dvdy.325 1162
37. Nojiri T, Wilson LAB, López-Aguirre C, Tu VT, Kuratani S, Ito K, et al. Embryonic evidence 1163
uncovers convergent origins of laryngeal echolocation in bats. Curr Biol [Internet]. 2021;31:1353-1164
1365.e3. Available from: http://dx.doi.org/10.1016/j.cub.2020.12.043 1165
38. Nojiri T, Tu VT, Sohn JH, Koyabu D. On the sequence heterochrony of cranial ossification of 1166
bats in light of Haeckel’s recapitulation theory. J Exp Zool B Mol Dev Evol [Internet]. 1167
2022;338:137–48. Available from: http://dx.doi.org/10.1002/jez.b.23042 1168
39. Sulser RB, Patterson BD, Urban DJ, Neander AI, Luo Z-X. Evolution of inner ear 1169
neuroanatomy of bats and implications for echolocation. Nature [Internet]. 2022;602:449–54. 1170
Available from: http://dx.doi.org/10.1038/s41586-021-04335-z 1171
40. Snipes CCG, Carter RT. The hyoid as a sound conducting apparatus in laryngeally 1172
echolocating bats. J Anat [Internet]. 2022;240:1020–33. Available from: 1173
http://dx.doi.org/10.1111/joa.13615 1174

41. Simmons NB, Geisler JH. Phylogenetic relationships of Icaronycteris, Archaeonycteris, 1175
Hassianycteris, and Palaeochiropteryx to extant bat lineages, with comments on the evolution of 1176
echolocation and foraging strategies in Microchiroptera. Bulletin of the AMNH; no. 235. 1998; 1177
42. Dobson GE. On the structure of the pharynx, larynx, and hyoid bones in the Epomophori; 1178
with remarks on its relation to the habits of these animals [Internet]. Proceedings of the Zoological 1179
Society of London; 1881. p. 685–93. Available from: 1180
https://www.biodiversitylibrary.org/item/46213 1181
43. Elias. Zur anatomie des Kehlkopfes der Mikrochiropteren [Internet]. Gegenbaurs 1182
morphologisches Jahrbuch; 1907. p. 736. Available from: 1183
https://www.biodiversitylibrary.org/item/137789 1184
44. Sprague JM. The hyoid region of placental mammals with especial reference to the bats. Am 1185
J Anat [Internet]. 1943;72:385–472. Available from: 1186
https://onlinelibrary.wiley.com/doi/10.1002/aja.1000720304 1187
45. Carter RT, Adams RA. Ontogeny of the larynx and flight ability in Jamaican fruit bats 1188
(Phyllostomidae) with considerations for the evolution of echolocation. Anat Rec [Internet]. 1189
2014;297:1270–7. Available from: http://dx.doi.org/10.1002/ar.22934 1190
46. Carter RT, Adams RA. Integrating ontogeny of echolocation and locomotion gives unique 1191
insights into the origin of bats. J Mamm Evol [Internet]. 2016;23:413–21. Available from: 1192
https://doi.org/10.1007/s10914-016-9324-2 1193
47. Carter RT. Reinforcement of the larynx and trachea in echolocating and non-echolocating bats. 1194
J Anat [Internet]. 2020;237:495–503. Available from: http://dx.doi.org/10.1111/joa.13204 1195
48. Snipes CCG, Carter RT. Vibroacoustic Response of the Tympanic Membrane to Hyoid-Borne 1196
Sound Generated during Echolocation in Bats. Integr Org Biol [Internet]. 2023;5:obad004. 1197
Available from: http://dx.doi.org/10.1093/iob/obad004 1198
49. Brualla NLM, Wilson LAB, Doube M, Carter RT, McElligott AG, Koyabu D. The vocal 1199
apparatus: An understudied tool to reconstruct the evolutionary history of echolocation in bats? J 1200
Mamm Evol. 2023;30:79–94. 1201
50. Metscher BD. MicroCT for developmental biology: a versatile tool for high-contrast 3D 1202
imaging at histological resolutions. Dev Dyn [Internet]. 2009;238:632–40. Available from: 1203
http://dx.doi.org/10.1002/dvdy.21857 1204

51. Gignac PM, Kley NJ, Clarke JA, Colbert MW, Morhardt AC, Cerio D, et al. Diffusible iodine-1205
based contrast-enhanced computed tomography (diceCT): an emerging tool for rapid, high-1206
resolution, 3-D imaging of metazoan soft tissues. J Anat [Internet]. 2016;228:889–909. Available 1207
from: http://dx.doi.org/10.1111/joa.12449 1208
52. Suthers RA, Fattu JM. Mechanisms of sound production by echolocating bats. Am Zool 1209
[Internet]. 1973 [cited 2023 Jan 4];13:1215–26. Available from: 1210
https://academic.oup.com/icb/article/13/4/1215/2094159 1211
53. Griffiths TA. Comparative laryngeal anatomy of the Big Brown bat, Eptesicus fuscus, and the 1212
Mustached bat, Pteronotus parnellii. Mammalia [Internet]. 1983 [cited 2023 Jan 6];47:377–94. 1213
Available from: https://www.degruyter.com/document/doi/10.1515/mamm-1983-0310/html 1214
54. Robin HA. Recherches anatomiques sur les mammifères de l’ordre des chiroptères [Internet]. 1215
Paris: Ann. Sci. Nat., Zool; 1881. Available from: 1216
https://www.biodiversitylibrary.org/item/106685 1217
55. Giannini NP, Wible JR, Simmons NB. On the cranial osteology of Chiroptera. I. pteropus 1218
(megachiroptera: Pteropodidae). Bull Am Mus Nat Hist [Internet]. 2006;295:1–134. Available 1219
from: http://www.bioone.org/doi/abs/10.1206/0003-1220
0090%282006%29295%5B0001%3AOTCOOC%5D2.0.CO%3B2 1221
56. de Bakker BS, de Bakker HM, Soerdjbalie-Maikoe V, Dikkers FG. The development of the 1222
human hyoid-larynx complex revisited. Laryngoscope [Internet]. 2018;128:1829–34. Available 1223
from: http://dx.doi.org/10.1002/lary.26987 1224
57. Rodríguez-Vázquez JF, Kim JH, Verdugo-López S, Murakami G, Cho KH, Asakawa S, et al. 1225
Human fetal hyoid body origin revisited. J Anat [Internet]. 2011;219:143–9. Available from: 1226
http://dx.doi.org/10.1111/j.1469-7580.2011.01387.x 1227
58. Henick DH. Three-dimensional analysis of murine laryngeal development. Ann Otol Rhinol 1228
Laryngol Suppl [Internet]. 1993;159:3–24. Available from: 1229
http://dx.doi.org/10.1177/00034894931020s301 1230
59. Alli O, Berzofsky C, Sharma S, Pitman MJ. Development of the rat larynx: a histological 1231
study. Laryngoscope [Internet]. 2013;123:3093–8. Available from: 1232
http://dx.doi.org/10.1002/lary.24145 1233
60. Hast MH. Early development of the human laryngeal muscles. Ann Otol Rhinol Laryngol 1234
[Internet]. 1972;81:524–30. Available from: http://dx.doi.org/10.1177/000348947208100411 1235

61. Förster G, Arnold D, Bischoff SJ, Schubert H, Scholle H-C, Müller AH. Laryngeal pacing in 1236
minipigs: in vivo test of a new minimal invasive transcricoidal electrode insertion method for 1237
functional electrical stimulation of the PCA. Eur Arch Otorhinolaryngol [Internet]. 1238
2013;270:225–31. Available from: http://dx.doi.org/10.1007/s00405-012-2141-1 1239
62. Cheetham J, Radcliffe CR, Ducharme NG, Sanders I, Mu L, Hermanson JW. Neuroanatomy 1240
of the equine dorsal cricoarytenoid muscle: surgical implications. Equine Vet J [Internet]. 1241
2008;40:70–5. Available from: http://dx.doi.org/10.2746/042516407X240465 1242
63. Mead AF, Osinalde N, Ørtenblad N, Nielsen J, Brewer J, Vellema M, et al. Fundamental 1243
constraints in synchronous muscle limit superfast motor control in vertebrates. Elife [Internet]. 1244
2017;6. Available from: http://dx.doi.org/10.7554/eLife.29425 1245
64. McKenna KZ, Wagner GP, Cooper KL. A developmental perspective of homology and 1246
evolutionary novelty. Curr Top Dev Biol [Internet]. 2021;141:1–38. Available from: 1247
http://dx.doi.org/10.1016/bs.ctdb.2020.12.001 1248
65. Galis F. Key innovations and radiations. In: Wagner GP, editor. The Character Concept in 1249
Evolutionary Biology. Academic Press; 2001. p. 581–605. 1250
66. Sekine K, Ohuchi H, Fujiwara M, Yamasaki M, Yoshizawa T, Sato T, et al. Fgf10 is essential 1251
for limb and lung formation. Nat Genet [Internet]. 1999;21:138–41. Available from: 1252
http://dx.doi.org/10.1038/5096 1253
67. Bachiller D, Klingensmith J, Shneyder N, Tran U, Anderson R, Rossant J, et al. The role of 1254
chordin/Bmp signals in mammalian pharyngeal development and DiGeorge syndrome. 1255
Development [Internet]. 2003;130:3567–78. Available from: http://dx.doi.org/10.1242/dev.00581 1256
68. Tabler JM, Rigney MM, Berman GJ, Gopalakrishnan S, Heude E, Al-Lami HA, et al. Cilia-1257
mediated Hedgehog signaling controls form and function in the mammalian larynx. Elife 1258
[Internet]. 2017;6. Available from: http://dx.doi.org/10.7554/eLife.19153 1259
69. Lungova V, Verheyden JM, Sun X, Thibeault SL. β-Catenin signaling is essential for 1260
mammalian larynx recanalization and the establishment of vocal fold progenitor cells. 1261
Development [Internet]. 2018;145. Available from: http://dx.doi.org/10.1242/dev.157677 1262
70. Adachi N, Bilio M, Baldini A, Kelly RG. Cardiopharyngeal mesoderm origins of 1263
musculoskeletal and connective tissues in the mammalian pharynx. Development [Internet]. 1264
2020;147. Available from: http://dx.doi.org/10.1242/dev.185256 1265

71. Novick A. Pulse Duration in the Echolocation of Insects by the Bat, Pteronotus. In: Autrum 1266
H, Bünning E, v. Frisch K, Hadorn E, Kühn A, Mayr E, et al., editors. Orientierung der Tiere / 1267
Animal Orientation: Symposium in Garmisch-Partenkirchen 17–21 9 1962 [Internet]. Berlin, 1268
Heidelberg: Springer Berlin Heidelberg; 1963. p. 21–6. Available from: 1269
https://doi.org/10.1007/978-3-642-99872-0_3 1270
72. Suthers RA, Hartley DJ, Wenstrup JJ. The acoustic role of tracheal chambers and nasal cavities 1271
in the production of sonar pulses by the horseshoe bat, Rhinolophus hildebrandti. J Comp Physiol 1272
A [Internet]. 1988;162:799–813. Available from: http://dx.doi.org/10.1007/BF00610969 1273
73. Rübsamen R, Schäfer M. Ontogenesis of auditory fovea representation in the inferior 1274
colliculus of the Sri Lankan rufous horseshoe bat, Rhinolophus rouxi. J Comp Physiol A [Internet]. 1275
1990;167:757–69. Available from: http://dx.doi.org/10.1007/BF00189766 1276
74. Funakoshi K, Nomura E, Matsukubo M, Wakita Y. Postnatal growth and vocalization 1277
development of the lesser horseshoe bat, Rhinolophus cornutus, in the Kyushu district, Japan. 1278
jmam [Internet]. 2010 [cited 2023 Jan 4];35:65–78. Available from: 1279
http://www.bioone.org/doi/full/10.3106/041.035.0105 1280
75. Morton ES. On the occurrence and significance of motivation-structural rules in some bird 1281
and mammal sounds. Am Nat. 1977;111:855–69. 1282
76. Ying LIU, Feng J, Jiang YL, Sun LWKP. Vocalization development of greater horseshoe bat, 1283
Rhinolophus ferrumequinum (Rhinolophidae, Chiroptera). Folia Zool Brno. 2007;56:126–36. 1284
77. Simmons NB, Seymour KL, Habersetzer J, Gunnell GF. Inferring echolocation in ancient bats 1285
[Internet]. Nature. 2010. p. E8; discussion E9. Available from: 1286
http://dx.doi.org/10.1038/nature09219 1287
78. Carter RT, Stuckey A, Adams RA. Ontogeny of the hyoid apparatus in Jamaican fruit bats 1288
(Chiroptera: Phyllostomidae) in unraveling the evolution of echolocation in bats. J Zool [Internet]. 1289
2019;308:301–7. Available from: https://onlinelibrary.wiley.com/doi/10.1111/jzo.12679 1290
79. Griffiths TA. Systematics of the New World nectar-feeding bats (Mammalia, Phyllostomidae), 1291
based on the morphology of the hyoid and lingual regions. American Museum novitates; no. 2742. 1292
1982; 1293
80. Griffiths TA, Criley BB. Comparative lingual anatomy of the bats Desmodus rotundus and 1294
Lonchophylla robusta (Chiroptera: Phyllostomidae). J Mammal [Internet]. 1989;70:608–13. 1295

Available from: http://www.jstor.org/stable/1381432 1296
81. Romankowowa A. Comparative study of the skeleton of the hyoid apparatus in some bat 1297
species; Badania porównawcze szkieletu aparatu gnykowego u niektórych gatunków nietoperzy. 1298
Acta Theriol [Internet]. 1963;7:15–23. Available from: 1299
https://rcin.org.pl/ibs/dlibra/publication/edition/9239 1300
82. Jebb D, Huang Z, Pippel M, Hughes GM, Lavrichenko K, Devanna P, et al. Six reference-1301
quality genomes reveal evolution of bat adaptations. Nature. 2020;583:578–84. 1302
83. Hedges SB, Kumar S. The timetree of life. OUP Oxford; 2009. 1303
84. Cretekos CJ, Weatherbee SD, Chen C-H, Badwaik NK, Niswander L, Behringer RR, et al. 1304
Embryonic staging system for the short-tailed fruit bat, Carollia perspicillata, a model organism 1305
for the mammalian order Chiroptera, based upon timed pregnancies in captive-bred animals. Dev 1306
Dyn [Internet]. 2005;233:721–38. Available from: http://dx.doi.org/10.1002/dvdy.20400 1307
85. Ito K, Tu VT, Eiting TP, Nojiri T. On the embryonic development of the nasal turbinals and 1308
their homology in bats. Frontiers in Cell and [Internet]. 2021; Available from: 1309
https://www.frontiersin.org/articles/10.3389/fcell.2021.613545/full 1310
86. Chi T-C, Tu VT, Sohn J, Kimura J, Koyabu D. Caudal auricular muscle variations and the 1311
evolution of echolocation behavior in pteropodid bats. J Vet Med Sci [Internet]. 2023;85:625–30. 1312
Available from: http://dx.doi.org/10.1292/jvms.23-0128 1313
87. Chi T-C, Meguro F, Takechi M, Furutera T, Tu VT, Higashiyama H, et al. Anatomy and 1314
homology of the caudal auricular muscles in greater short-nosed fruit bat (Cynopterus sphinx). J 1315
Vet Med Sci [Internet]. 2023;85:571–7. Available from: http://dx.doi.org/10.1292/jvms.23-0088 1316
88. Sohn JH, Fukui D, Nojiri T, Minowa K, Kimura J. Three-Dimensional and Histological 1317
Observations on Male Genital Organs of Greater Horseshoe Bat, Rhinolophus ferrumequinum. J 1318
Mamm Evol [Internet]. 2021; Available from: https://link.springer.com/article/10.1007/s10914-1319
020-09525-6 1320
89. Jungers WL, Falsetti AB, Wall CE. Shape, relative size, and size-adjustments in 1321
morphometrics. Am J Phys Anthropol [Internet]. 1995;38:137–61. Available from: 1322
https://onlinelibrary.wiley.com/doi/10.1002/ajpa.1330380608 1323
90. Koyabu D, Werneburg I, Morimoto N, Zollikofer CPE, Forasiepi AM, Endo H, et al. 1324
Mammalian skull heterochrony reveals modular evolution and a link between cranial 1325

development and brain size. Nat Commun [Internet]. 2014;5:3625. Available from: 1326
http://dx.doi.org/10.1038/ncomms4625 1327
91. Aiello, WG, Freedman, HI, Wu, J. Analysis of a model representing stage-structured 1328
population growth with state-dependent time delay. SIAM J Appl Math. 1992;52:855–69. 1329
92. Rayner JMV. Linear relations in biomechanics: the statistics of scaling functions. J Zool. 1330
1985;206:415–39. 1331
93. Nelder JA, Wedderburn RWM. Generalized Linear Models. J R Stat Soc Ser A [Internet]. 1332
1972;135:370. Available from: https://www.jstor.org/stable/10.2307/2344614?origin=crossref 1333