A functional comparison of the hyolingual complex in pygmy and dwarf sperm whales (Kogia breviceps and K. sima), and bottlenose dolphins (Tursiops truncatus)

Abstract

The function of the hyolingual complex in three odontocete species was investigated to compare adaptations of divergent feeding strategies, suction and ram feeding. Pygmy and dwarf sperm whales, members of the genus Kogia (or kogiids), are known to be strong suction feeders. We tested the hypotheses that kogiids would have a larger, more robust hyolingual complex, and that hyolingual muscles would have a greater maximum theoretical muscle tension compared with ram-based feeders such as bottlenose dolphins (Tursiops truncatus). A robustness index and surface area was calculated for bony hyoid elements in kogiids and bottlenose dolphins. The anatomy, muscle architecture, pinnation, two-dimensional angle of attachment and maximum theoretical muscle tension were measured in each hyolingual muscle. A functional model incorporating vector analyses of hyolingual musculature was created for kogiids and bottlenose dolphins to assess differences in function of their hyolingual complexes. Kogiid hyoid surface areas were significantly greater (P = 0.01) than in bottlenose dolphins. Most maximum theoretical muscle tensions of hyolingual complexes were not significantly different within or between species. The data suggest that associated orofacial and tongue morphology, particularly the relationship of hyoid shape and tongue retractor muscles, greatly influences suction capability in odontocetes. Kogiids demonstrated adaptations that occlude lateral gape, including a novel vertical ridge on each side of the mandible, and a shortened mandible that is capable of a large gape, and gape angle. These adaptations presumably assist in maintaining negative intraoral pressure generated by the depression and retraction of the relatively short and wide kogiid tongue. The tongues of kogiids should be capable of generating greater intraoral volume changes compared with the long, narrow tongue of bottlenose dolphins.

Introduction

Odontocetes (toothed whales) display a broad repertoire of feeding behaviours, which are due to a tremendous diversity of feeding apparatus morphology, and associated structures (reviewed by Werth, 2000a; Marshall, 2002). Despite this diversity, odontocetes are frequently over-generalized as long-snouted taxa that capture prey by chasing and grabbing their prey with dozens or hundreds of teeth (Werth, 1992). Although many delphinids such as bottlenose dolphins (Tursiops truncatus) exhibit this morphology (Slijper, 1962; Norris & Møhl, 1983; Heyning, 1989; Rommel, 1990), not all odontocetes do. In fact, the greatest breadth of morphological diversity is found among the non-delphinid odontocetes (Gaskin, 1976; Werth, 2000a). The use of suction to feed is probably common among this group, but has only been demonstrated for a few species (beaked whales, Heyning & Mead, 1996; pilot whales, Werth, 2000b; harbor porpoise, Kastelein et al. 1997a; pygmy and dwarf sperm whales, Bloodworth & Marshall, 2005). Among marine mammals, the underlying mechanism for generating suction is presumably the rapid depression and retraction of the hyolingual complex (lingual musculature and hyoid apparatus), which increases oral cavity volume and reduces intraoral pressure (e.g. Gordon, 1984; Werth, 2000a). Morphological changes, such as enlarged hyoids and increased hyolingual muscle size, have been hypothesized to be related to generating powerful suction forces (Werth, 1992; Heyning & Mead, 1996). Indeed, enlarged hyolingual musculature has been identified in short-finned pilot whales, ziphiids, physeterids and kogiids, all of whom are presumed suction feeders (Reidenberg & Laitman, 1994; Werth, 1992, 2004; Heyning & Mead, 1996). Hyolingual morphology and the presence of throat grooves in members of the genus Kogia (family Kogiidae; kogiids) suggest a powerful suction capability that could be among the best developed among odontocetes (Caldwell & Caldwell, 1989; Werth, 1992; Reidenberg & Laitman, 1994). Odontocete suction feeding has been associated with blunt rostra, few, if any, functional teeth, short wide mandibles, enlarged hyoids and occasionally throat grooves (Ross, 1978; Reidenberg & Laitman, 1994; Heyning & Mead, 1996; Werth, 2000b, 2006; Bloodworth & Marshall, 2005). Deep-diving ability and a teuthophagous diet are behavioural traits also associated with odontocetes that presumably use suction (Ross, 1978; Pinedo, 1987; Dos Santos & Haimovici, 2001; Wang et al. 2002).

Pygmy sperm whales (Kogia breviceps) and dwarf sperm whales (K. sima) are of particular interest due to their apparently increased suction capability over other odontocetes, their high stranding frequency in the south-east US and their distant phylogenetic relationship to delphinids, such as bottlenose dolphins (Odell, 1991; Milinkovitch et al. 1994; Berta & Sumich, 1999; Geisler & Sanders, 2003). In a prior comparative kinematic study of kogiids and bottlenose dolphins (Bloodworth & Marshall, 2005), kogiids were confirmed to use suction as their primary feeding mode. Bottlenose dolphins were verified as ram-based feeders with little, if any, suction behaviour observed in controlled feeding trials. Most delphinids with elongated rostra are probably incapable of significant suction feeding because negative intraoral pressures cannot be maintained over the relatively long distances of the snout. However, comparative studies of presumed suction and ram-based feeding in odontocetes are sparse, and have not adequately described functional differences that support various feeding strategies. The aim of this study was to characterize the functional morphology of the hyolingual apparatus of widely dissimilar odontocetes, kogiids and bottlenose dolphins. Investigations of the hyolingual complex of all three species investigated the following hypotheses: (1) kogiid hyolingual muscles are capable of significantly greater maximum theoretical muscle tensions than bottlenose dolphins, and (2) bones of the hyoid apparatus are more robust and exhibit greater surface areas in kogiids than bottlenose dolphins.

Materials and methods

Salvage collection

Salvage material included five K. sima, four K. breviceps and six T. truncatus heads collected from freshly dead or moderately decomposed stranded specimens with the assistance of the Southeast U.S. Marine Mammal Stranding Network, Texas Marine Mammal Stranding Network (TMMSN) and Mote Marine Laboratory (MML) under NOAA Salvage Permit Number SER02-259 to C.D.M. Heads were usually decapitated, transecting the sternohyoideus and sternothyroideus muscles. However, when possible, all muscles associated with the hyoid were retained by removing the head with the sternum.

Hyoid osteology

To test whether kogiid basihyal and thyrohyal surface areas were greater than in bottlenose dolphins, hyoids were removed, cleaned of tissue and dried for morphometric analysis. Ventral basihyal and thyrohyal surfaces were digitally photographed, and their surface areas calculated using Image J (NIH, Bethesda, MD, USA). Stylohyals were similarly cleaned and dried. Lengths and maximum widths at each stylohyal's mid-length were measured with vernier calipers to calculate a stylohyal robustness index (width/length). Tympanohyal, stylohyal and epihyal arcs of rotation relative to the palatal plane were measured from digital images of a cleaned head of each species with the hyoid at rest, and in fully depressed and retracted (hyoid manually pulled caudally to maximum extent) positions.

Hyolingual myology

Dissections were conducted at the TMMSN's Necropsy Laboratory in Galveston, Texas and at Texas A&M University at Galveston. Heads were thawed at room temperature until the blubber softened, at which time orofacial morphology was noted and photographed. The origins and insertions of hyal and extrinsic lingual muscles were identified. Muscle nomenclature followed ICVGAN (2005) except when a unique muscle was observed, in which case classic cetacean literature was referenced. The sternothyroideus muscle was included as a potential hyoid retractor. Two-dimensional angles of muscle insertion relative to the palatal plane were measured in the parasagittal plane with a protractor in situ. A list of abbreviations can be found at the end of the Methods section.

Muscle characterization

Physiological cross-sectional area (PCSA) was obtained to calculate maximum theoretical muscle tension (MTMT) for each hyolingual muscle. Physiological cross-sectional area was estimated following Spector et al. (1980) and Powell et al. (1984):

$$\text{PCSA}=\left(\text{muscle mass}\right)\times (\text{cos}\Theta )/\left(\text{FL}\right)\times \left(\text{muscle density}\right)$$

where cos θ was mean fascicle angle relative to a muscle's line of action, FL was mean fibre length (cm) and muscle density was assumed to be 1.056 g cm⁻³ (Mendez & Key, 1960; Murphy & Beardsley, 1974). After several unsuccessful attempts to measure muscle fibre lengths from salvage material, fibre lengths were assumed to equal fascicle lengths, as all muscles were short and parallel-fibred and fascicle length approximates fibre length in these types of muscles (Alexander, 1968; Gans & De Vree, 1987). This assumption was tested using anatomical cross-sectional data of each muscle, which provided results similar to PCSA data. Maximum theoretical muscle tension was estimated following Spector et al. (1980), Roy et al. (1982) and Powell et al. (1984):

$$\text{MTMT}=\left(\text{PCSA}\right)\times \left(\text{specific tension}\right)$$

where specific tension was assumed to be 22.5 N cm⁻². Angles relative to the palatal plane and MTMT were used to calculate muscle vectors and create two-dimensional models of hyolingual function in kogiids and bottlenose dolphins. Hyoid depression and retraction vectors were calculated to determine predicted net tensions exerted upon the entire hyoid complex at the moment of contraction. Similar vector predictions were made for lingual depressors and retractors.

Statistical analyses

Osteological and myological data were grouped by species and genera for analyses. Comparisons included K. sima to K. breviceps, and both species of Kogia to T. truncatus. To control for scaling effects of body size, an analysis of covariance (ancova; SPSS Statistical Software package 11.0) was used to identify significant differences in PCSA and MTMT using total length or mass as a covariate. Because mass was a less reliable predictor, variable means were scaled to a total body length of the greatest r²value, which provided the most accurate basis for ancova testing. A one-way analysis of variance (anova) was used to test for differences in right vs. left muscle and hyoid variables. Where three or more MTMT data points were available, normality was met in all groups, with the exceptions of pooled Kogia styloglossus muscle (0.030) and T. truncatus geniohyoideus muscle (0.012).

Abbreviations

CT, cricothyroideus; GG, genioglossus; GH, geniohyoideus; HG, hyoglossus; IH, interhyoideus; MH, mylohyoideus; OO, orbicularis oris; SC, sternocephalicus; ST, sternothyroideus; SteH, sternohyoideus; StyG, styloglossus; StyH, stylohyoideus; TH, thyrohyoideus.

Results

Hyoid osteology

The hyoid apparatuses of kogiids and bottlenose dolphins were composed of the same components present in terrestrial mammals: tympanohyal, stylohyal, ceratohyal, epihyal, basihyal and thyrohyal elements (Figs 1 and 2). Overall, the hyoid progressed rostroventrally until the epihyals turned caudoventrally to the basihyal. In kogiids, the first three segments, the tympanohyal, stylohyal and ceratohyal, comprised a fused rostroventrally and mediolaterally directed chain. In bottlenose dolphins, ossification was absent in the tympanohyal, epihyal and ceratohyal, which all appeared relatively reduced in size. Overall, tympanohyal, epihyal and ceratohyal morphologies were similar to kogiids, but were smaller and more gracile. In both kogiids and bottlenose dolphins, the thyrohyals branched caudolaterally from the basihyal, and were fused to it in adult specimens, forming a broad surface area for muscle attachment. In immature kogiid specimens, the tympanohyals and epihyals consisted of large, cartilaginous rods. In older kogiid specimens, both elements were reduced to small, cartilaginous plates between the stylohyals and tympanic bones (tympanohyal) or ceratohyals (epihyal). Manipulation of kogiid hyolingual complexes demonstrated restricted caudoventral hyoid movement in K. sima due to a thyrohyal–thyroid ligament that severely limited hyoid mobility. As a result, the basihyal and thyrohyals in K. sima were incapable of abduction away from the larynx. Significant separation of the fused basihyal–thyrohyal complex from the stylohyal was prevented by a taut connection of these elements by the interhyoideus muscle, but up to 30° of rotation relative to the palatal plane was observed in both kogiid and bottlenose dolphin hyoids during manipulation (Fig. 2).

Fig. 1.

Ventral views of the ossified hyoid complex elements. (A) K. breviceps. (B) T. truncatus. Note prominences for epihyal articulations along the rostromedial basihyal surface.

Fig. 2.

Skulls, mandibles and ossified hyoid elements. (A) K. breviceps and (B) T. truncatus viewed from the ventral aspect. (C) K. breviceps and (D) T. truncatus viewed from a lateral perspective with hyoids in representative retracted positions.

Mean basihyal surface areas (at a scaled total body length of 227 cm) were 35.6 cm² for K. sima and 36.5 cm² for K. breviceps. Mean scaled thyrohyal surface areas were 32.3 cm² (right) and 30.6 cm² (left) for K. sima vs. 25.6 cm² (right) and 26.6 cm² (left) for K. breviceps. Overall, mean scaled basihyal–thyrohyal surface areas were 98.6 cm² for K. sima and 88.7 cm² for K. breviceps. The mean bottlenose dolphin basihyal surface area at a scaled total body length of 227 cm was 15.2 cm². The mean scaled thyrohyal surface areas were 9.16 cm² (right) and 8.84 cm² (left) and the scaled combined basihyal–thyrohyal surface area was 33.2 cm². Mean scaled hyoid osteology variables are summarized in Table 1.

Table 1.

Scaled means (± SE) of hyoid osteology variables for K. sima (K.s.; d.f. = 4), K. breviceps (K.b.; d.f. = 3), pooled Kogia (Kog; d.f. = 8) and T. truncatus (T.t.; d.f. = 5) at a standardized body length of 227 cm

	Surface area (cm2)				Width (cm)		Length (cm)		Width/length
Species	Basihyal	Right thyrohyal	Left thyrohyal	Total basi/thyrohyal	Right stylohyal	Left stylohyal	Right stylohyal	Left stylohyal	Right stylohyal	Left stylohyal
K.s.	35.6 ± 2.58	32.3 ± 3.15	30.6 ± 3.19	98.6 ± 8.47	9.45 ± 0.71	10.5 ± 0.58	74.2 ± 3.91	75.0 ± 3.58	0.127 ± 0.01	0.14 ± 0.01
K.b.	36.5 ± 3.31	25.6 ± 4.05	26.6 ± 4.09	88.7 ± 10.9	9.45 ± 0.71	9.69 ± 0.74	80.2 ± 5.02	76.7 ± 4.59	0.118 ± 0.01	0.13 ± 0.01
Kog	36.0 ± 2.02	29.8 ± 2.47	29.1 ± 2.50	94.9 ± 6.63	10.3 ± 0.44	10.2 ± 0.45	76.5 ± 3.06	75.6 ± 2.80	0.135 ± 0.01	0.135 ± 0.01
T.t.	15.2 ± 2.35	9.16 ± 2.87	8.84 ± 2.90	33.2 ± 7.71	8.66 ± 0.51	8.71 ± 0.53	84.5 ± 3.56	84.5 ± 3.26	0.102 ± 0.01	0.103 ± 0.01

All Kogia groups had significantly greater combined basihyal–thyrohyal surface areas than bottlenose dolphins (P = 0.001; adjusted r² = 0.830). All Kogia groups had significantly greater basihyal (P = 0.001; adjusted r² = 0.892) and thyrohyal (P = 0.01; adjusted r² = 0.780) surface areas than bottlenose dolphins. Significant differences were not found between K. sima and K. breviceps hyoid variables, and no asymmetries were detected. Kogiid and bottlenose dolphin mean scaled stylohyal lengths (which can be considered lever arms that could change mechanical advantage) were not significantly different. Stylohyal robustness indices were not significantly different between groups, and also did not show any asymmetries.

Comparative hyolingual myology

The hyolingual myology of both kogiids and bottlenose dolphins was characterized by bilateral parallel-fibred muscles that, with the exception of lingual insertions, attached by aponeuroses to hyal, laryngeal and sternal elements. The stylohyoideus and cricothyroideus muscles were not observed in either kogiid species, and were assumed to be absent. The stylohyoideus and cricothyroideus muscles were identified in bottlenose dolphins. However, the buccinator was not found and is generally less prominent in cetaceans than in terrestrial mammals. Mean scaled kogiid and bottlenose muscle variables are summarized in Table 2.

Table 2.

Scaled means (± SE) of hyolingual myology length, mass, physiological cross-sectional area (PCSA), maximum theoretical muscle tension (MTMT), and parasagittal angle for K. sima (K.s.), K. breviceps (K.b.), pooled Kogia (Kog) and T. truncatus (T.t.) at a standardized body length of 252 cm. Cricothyroideus and stylohyoideus values are set to 241 cm for greatest accuracy in T.t. scaling

	Species (d.f.)	Length (mm) (d.f.)	Mass (g) (d.f.)	PCSA (cm2) (d.f.)	MTMT (N) (d.f.)	Parasagittal angle (d.f.)
SteH	K.s. (2)	300 ± 24.3	407 ± 36.0	12.8 ± 1.16	288 ± 26.1	160 ± 4.02
	K.b. (1)	292 ± 42.9	365 ± 63.7	11.3 ± 2.05	253 ± 46.2	166 ± 7.38
	Kog (4)	297 ± 19.2	393 ± 28.6	12.3 ± 0.92	276 ± 20.7	162 ± 3.30
	T.t. (4)	246 ± 20.5	346 ± 30.5	12.8 ± 0.98	288 ± 22.1	153 ± 2.69
MH	K.s.	107 ± 12.6 (3)	93.3 ± 11.1 (4)	8.00 ± 1.69 (3)	180 ± 38.1 (3)	–69.7 ± 9.81 (4)
	K.b.	117 ± 22.3 (1)	82.0 ± 19.6 (1)	5.66 ± 3.00 (1)	127 ± 67.4 (1)	–67.6 ± 17.3 (2)
	Kog	111 ± 10.0 (5)	89.6 ± 8.78 (6)	7.22 ± 1.34 (5)	162 ± 30.2 (5)	–69.0 ± 7.77 (7)
	T.t.	101 ± 10.7 (4)	128 ± 9.38 (3)	12.0 ± 1.43 (3)	270 ± 32.2 (3)	–79.3 ± 8.30 (4)
ST	K.s.	187 ± 11.3 (3)	79.3 ± 1.02 (3)	4.12 ± 0.34 (4)	92.7 ± 7.66 (4)	148 ± 16.6 (1)
	K.b. (1)	182 ± 20.0	78.7 ± 1.81	3.98 ± 0.60	89.5 ± 13.5	152 ± 30.5
	Kog	185 ± 9.0 (5)	79.1 ± 0.81 (5)	4.07 ± 0.27 (6)	91.6 ± 6.07 (6)	149 ± 13.6 (3)
	T.t. (4)	178 ± 9.57	34.4 ± 0.86	1.66 ± 0.29	37.4 ± 6.48	134 ± 11.1
GH	K.s. (4)	138 ± 35.3	31.3 ± 6.91	2.05 ± 0.60	46.1 ± 13.6	–14.5 ± 7.55
	K.b. (2)	187 ± 62.5	29.5 ± 12.2	1.38 ± 1.07	31.0 ± 24.0	–3.20 ± 13.9
	Kog (7)	154 ± 28.0	30.7 ± 5.47	1.82 ± 0.48	41.0 ± 10.7	–10.7 ± 6.19
	T.t. (5)	209 ± 29.9	49.7 ± 5.84	2.43 ± 0.51	54.6 ± 11.5	–26.4 ± 5.05
HG	K.s.	130 ± 36.8 (3)	11.0 ± 7.21 (4)	0.67 ± 0.31 (3)	15.2 ± 6.92 (3)	149 ± 10.6 (3)
	K.b. (1)	102 ± 65.2	1.26 ± 12.8	0.60 ± 0.54	13.5 ± 12.2	142 ± 19.5
	Kog	121 ± 29.2 (5)	7.77 ± 5.71 (6)	0.65 ± 0.24 (5)	14.6 ± 5.48 (5)	147 ± 8.71 (5)
	T.t. (5)	199 ± 31.2	55.2 ± 6.10	2.69 ± 0.26	60.5 ± 5.86	149 ± 7.11
TH	K.s.(4)	104 ± 3.28	11.8 ± 1.32	1.05 ± 0.18	23.5 ± 3.99	–161 ± 13.0
	K.b. (2)	105 ± 5.81	13.5 ± 2.33	1.22 ± 0.31	27.4 ± 7.05	–161 ± 23.9
	Kog (7)	104 ± 2.60	12.4 ± 1.04	1.10 ± 0.14	24.8 ± 3.16	–161 ± 10.7
	T.t. (5)	79.6 ± 2.78	10.2 ± 1.11	1.24 ± 0.15	27.9 ± 3.37	–147 ± 8.73
IH	K.s. (4)	18.6 ± 1.0	50.8 ± 2.25	1.30 ± 0.34	29.2 ± 7.62	–122 ± 13.7
	K.b. (1)	17.7 ± 1.77	50.8 ± 3.97	1.63 ± 0.60	36.6 ± 13.5	–134 ± 24.2
	Kog (6)	18.3 ± 0.79	50.8 ± 1.78	1.41 ± 0.27	31.7 ± 6.04	–126 ± 10.8
	T.t. (3)	11.6 ± 0.84	27.4 ± 1.90	0.90 ± 0.29	20.2 ± 6.45	–142 ± 11.6
GG	K.s.(4)	156 ± 36.9	29.4 ± 4.06	1.68 ± 0.19	37.7 ± 4.19	–8.30 ± 5.10
	K.b. (2)	205 ± 65.2	33.8 ± 7.19	1.54 ± 0.33	34.7 ± 7.40	–9.30 ± 9.36
	Kog (7)	172 ± 29.2	30.9 ± 3.22	1.63 ± 0.15	36.7 ± 3.32	–8.60 ± 4.18
	T.t. (5)	228 ± 31.2	32.5 ± 3.44	1.32 ± 0.16	29.7 ± 3.54	–8.30 ± 3.41
StyG	K.s.(4)	170 ± 19.7	12.5 ± 3.56	0.66 ± 0.10	14.9 ± 2.29	167 ± 3.54
	K.b. (2)	195 ± 34.9	15.5 ± 6.30	0.69 ± 0.18	15.6 ± 4.04	166 ± 6.49
	Kog (7)	178 ± 15.6	13.5 ± 2.82	0.67 ± 0.08	15.2 ± 1.81	166 ± 2.90
	T.t. (5)	220 ± 16.7	28.4 ± 3.01	1.25 ± 0.09	28.1 ± 1.93	170 ± 2.37
StyH	T.t. (5)	29.0 ± 1.78	1.47 ± 0.17	0.67 ± 0.02	15.0 ± 0.40	–64.1 ± 0.96
CH	T.t. (4)	60.8 ± 5.9	5.23 ± 1.54	0.80 ± 0.17	18.0 ± 3.84	145 ± 4.93

The sternohyoideus muscle (Figs 3A,C and 4A,C) originated upon the medial aspect of the manubrium's cranioventral surface, and the ventromedial surface of the first true rib in bottlenose dolphins. The sternohyoideus muscle inserted on nearly the entire ventral basihyal–thyrohyal surface (with the exception of the thyrohyoideus insertion), leaving only a narrow border available for attachment of several hyolingual muscles rostral to the basihyal and thyrohyals. The sternohyoideus probably functioned to draw the hyoid caudoventrally, had the greatest MTMT values of any kogiid or bottlenose dolphin hyolingual muscle and was probably the primary hyoid depressor and retractor.

Fig. 3.

Kogiid hyolingual muscles. (A) Ventral view of caudal hyolingual musculature. (B) Ventral view of rostral hyolingual musculature. (C) Hyolingual musculature viewed laterally. Note: the geniohyoideus is cut to clarify its insertion to the sternohyoideus. The hyoglossus is also partly reflected to reveal the interhyoideus.

Fig. 4.

Bottlenose dolphin hyolingual muscles. (A) Ventral view of caudal hyolingual musculature. (B) Ventral view of rostral hyolingual musculature. (C) Hyolingual musculature viewed laterally.

In both kogiids and bottlenose dolphins, the mylohyoideus muscle (Figs 3 and 4) was a broad, relatively thin muscle that originated upon the ventromedial mandibular border of both rami from the mandibular symphysis to near the mandible's caudal extent. Right and left muscles inserted to a ventromedial raphe that formed a lingual sling. The insertion extended from the mandibular symphysis to near the rostral basihyal border, where a fascial sheet joined right and left muscles. In bottlenose dolphins, a ‘V’-shaped fascial sheet joined each side at the mediocaudal aspect, but only extended approximately one-fifth of the mylohyoideus width from the basihyal. Although thin sagittally (2–3 mm), the mylohyoideus thickened (2–3 cm) laterally near its origin. The mylohyoideus probably functioned to return the hyoid to its basal position. The mylohyoideus was superficial to the platysma and digastricus muscles, while the deep stylohyoideus, styloglossus, hyoglossus and genioglossus muscles were loosely attached to the mylohyoideus by fascia. Maximum theoretical muscle tension values of the mylohyoideus were the second highest of all muscles studied in both kogiids and bottlenose dolphins (Table 2).

The sternothyroideus muscle (Figs 3A and 4A) originated upon the craniolateral surface of the ventral manubrium in bottlenose dolphins, and over the articulation of the manubrium with the first true rib in kogiids. This muscle extended craniodorsally to insert upon the caudolateral and rostrolateral surfaces of the thyroid cartilage and its rostral cornu. Maximum theoretical muscle tension values of the sternothyroideus were the third highest for kogiids but only the fifth highest in bottlenose dolphins. The sternothyroideus was probably the primary retractor and depressor of the larynx.

The geniohyoideus muscle (Figs 3A,C and 4B) originated upon the caudoventral surface of the mandibular symphysis by a long tendon shared by right and left muscles. The tendon continued approximately one-third of the distance to the insertion, where a sharp fascicular transition occurred. Although fascicles intermingled caudally, a distinct division was present between right and left muscles. This muscle inserted on the rostromedial sternohyoideus border in kogiids, but inserted on the rostromedial basihyal edge, just rostral to the sternohyoideus, rather than on the sternohyoideus itself in bottlenose dolphins. It is probable that the geniohyoideus assisted mandibular depression when the hyoid was depressed and, after mandibular elevation, assisted hyoid return. Maximal theoretical muscle tension was the fourth highest in kogiids and bottlenose dolphins. The geniohyoideus was deep to the mylohyoideus, and superficial to the genioglossus muscle along the entire genioglossal length, as well to the deep hyoglossus muscle near its insertion.

The hyoglossus muscle (Fig. 3B,C) originated upon the rostral basihyal border (in kogiids) or the rostral basihyal and thyrohyal borders (in bottlenose dolphins) as a thin, triangular aponeurosis from the basihyal–ceratohyal articulation, laterally to the caudolateral thyrohyal extremity. In kogiids, the hyoglossus narrowed to insert diffusely in the middle one-third of the tongue body in K. sima, and in the caudal one-third in K. breviceps. In bottlenose dolphins, this muscle became wider as it extended rostrodorsally to insert diffusely into the caudal one-third to two-thirds of the ventrolateral tongue body. The hyoglossus had the third highest maximum MTMT in bottlenose dolphins, but was the least powerful muscle in kogiids. The hyoglossus pulled the tongue caudoventrally, and was probably the primary lingual depressor.

In kogiids, the thyrohyoideus muscle (Fig. 3A) originated upon the ventrolateral thyroid cartilage and the rostral cornu of the arytenoid cartilage. However, in bottlenose dolphins (Fig. 4A,B), it originated from the craniolateral surface of the rostral cornu and surface of the laryngeal thyroid cartilage. This muscle extended rostromedially to insert upon the medial basihyal and thyrohyals, caudal to the insertion of the sternohyoideus muscle in kogiids, and the ventrocaudal and medial aspects of the basihyal in bottlenose dolphins. The function of the thyrohyoideus was to draw the hyoid caudodorsally. The thyrohyoideus was deep to the sternohyoideus and was bordered caudodorsally by the sternothyroideus.

The interhyoideus muscle (Figs 3C and 4C; same as interhyoid by Lawrence & Schevill, 1965) was a short, thick and wide muscle that originated upon the ventral stylohyal and caudal ceratohyal surfaces. It inserted upon the entire dorsal basihyal and thyrohyal surfaces; all other hyoid muscles attached to the ventral, rostral or lateral hyal surfaces. The interhyoideus functioned to prevent significant separation of the basihyal and thyrohyal from the stylohyal, and may have approximated these attachments.

The genioglossus muscle (Figs 3B,C and 4B,C) originated upon the ventral mandibular symphysis in bottlenose dolphins and the adjacent medial mandibular surface in kogiids. The insertion in K. sima was dispersed in the tongue's entire ventral body, while in K. breviceps, the insertion was centralized into the middle one-third of the tongue's ventrolateral body. In bottlenose dolphins, the insertion was variable and ranged from the caudal one-third of the ventromedial tongue body to the entire ventromedial tongue surface. The genioglossus was the only muscle capable of tongue protraction.

The styloglossus muscle (Figs 3C and 4C) originated from the stylohyal's dorsocaudal surface and inserted rostrally along the middle one-third of the lateral tongue body. Although small, the styloglossus was probably the primary tongue retractor and drew the tongue caudoventrally. The styloglossus was deep to the mylohyoideus.

The stylohyoideus muscle (Fig. 4C) was found only in bottlenose dolphins. It originated on the ventrocaudal and caudolateral stylohyal surfaces, and narrowed to insert by a tendon to the caudolateral tip of the thyrohyal. The stylohyoideus muscle probably anchored the thyrohyal to the stylohyal. Although small, the stylohyoideus was distinct and was located caudal to the rostral mylohyoideus and hyoglossus muscles.

The cricothyroideus muscle (Fig. 4A) was also identified only in bottlenose dolphins. It originated upon the caudoventral and caudolateral border of the cricoid cartilage of the larynx. This small muscle inserted upon the caudal border of the rostral cornu of the thyroid cartilage. The cricothyroideus approximated the cricoid and thyroid cartilages and was deep to the sternohyoideus.

Maximum theoretical muscle tension

It was hypothesized that kogiid muscles were capable of significantly greater MTMT than bottlenose dolphins, which would produce more powerful hyolingual depression and retraction. However, ancovas of MTMT values rejected this hypothesis and identified only a few significant differences between specific muscles of kogiids and bottlenose dolphins after scaling for either total body length or mass. Styloglossus MTMT values were significantly less in K. sima (P < 0.005) and pooled Kogia (P < 0.01) than in T. truncatus (adjusted r² = 0.732). Hyoglossus MTMT values were significantly less in all Kogia groups (P < 0.05) than in T. truncatus (adjusted r² = 0.758). Sternothyroideus MTMT values were significantly greater in K. sima (P < 0.01) and pooled Kogia (P < 0.05) than in T. truncatus (adjusted r² = 0.656). Asymmetry was not significant within or between any species.

Functional model of hyolingual function in kogiids and bottlenose dolphins

A model of hyoid depression and retraction in kogiids and bottlenose dolphins (Figs 5 and 6) demonstrated that the hyoid complex of kogiids is predicted to contract caudoventrally with a force of 771 N at an angle of –18.7°. In comparison, the hyoid complex of bottlenose dolphins is predicted to contract caudoventrally with a force of 675 N at an angle of –25.0°. Specifically, the tongue of kogiids is predicted to be pulled caudoventrally at a force of 58.8 N at an angle of –23.3°, while the tongue of bottlenose dolphins is predicted to be pulled caudoventrally with a force of 174.4 N at an angle of –24.7°. The sternohyoideus is the major depressor–retractor of the hyoid complex, and the hyoglossus is the major depressor–retractor of the tongue for both kogiids and bottlenose dolphins. These values indicate that bottlenose dolphins are able to retract the tongue approximately three times more forcefully than kogiids, but at approximately similar angles. The hyoglossus is the major depressor and retractor of the bottlenose dolphin's tongue, while this role is shared by the styloglossus and the hyoglossus muscles in kogiids.

Fig. 5.

Functional models of the hyolingual complex in kogiids. Forces reflect combined right and left muscles. (A) Black lines are lines of action for hyoid muscles. Grey lines are vector resultants for hyal depressors and retractors. Dashed lines represent the outline of the stylohyal. (B) Schematic of kogiid lingual muscles and tongue. Black lines are lines of action for lingual muscles. Grey lines are vector resultants for lingual depressors and retractors.

Fig. 6.

Functional models of the hyolingual complex in bottlenose dolphins. Forces reflect combined right and left muscles. (A) Black lines are lines of action for hyoid muscles. Grey lines are vector resultants for hyal depressors and retractors. (B) Schematic of bottlenose dolphin lingual muscles and tongue. Black lines are lines of action for lingual muscles. Grey lines are vector resultants for lingual depressors and retractors.

Associated kogiid and bottlenose dolphin orofacial morphology

Kogiid orofacial morphology, including the tongue, is probably critical to pygmy and dwarf sperm whales for generating suction. The kogiid tongue was triangular, with a wide caudal portion, and was contoured to fit into a raised palatal vault. Rostral manipulation of the tongue was restricted by muscular hyal connections, but caudal manipulation was easily facilitated by hyoid depression or retraction. The blubber rostral to the hyoid was more fibrous and less pliable, but caudally there was a sharp transition to more elastic blubber. Lateral gape (lateral sides of the mouth) was partially occluded caudally by the buccinator, and orbicularis oris muscles, which were overlaid by blubber and skin. In addition, a dense fibrous ridge of tissue was present on each lateral margin of the jaw and formed a vertical wall that further occluded lateral gape when the jaw was depressed. This elastic structure fit into a recess of the blubber laterally when the mouth was closed. Only the rostral-most mandibular teeth were visible laterally when the mouth was open due to the presence of this structure; the majority of lateral gape was occluded.

The tongue of bottlenose dolphins was long and narrow compared with the wide, pointed tongue of kogiids. As in kogiids, the tongue of bottlenose dolphins fit into a palatal elevation, but the vault was relatively shallow compared with the deep palatal vault of kogiids. The gular blubber in bottlenose dolphins appeared to shift gradually from a pliable state over the trunk to a progressively less elastic character near the hyoid's rostral border. Unlike kogiids there were no structures that limited lateral gape.

Discussion

Heyning & Mead (1996) considered hyoid surface area and enlarged hyolingual musculature to be adaptations for suction feeding in beaked whales. Their reasoning was that larger, more powerful muscles are capable of producing greater suction. Recently, it was verified that kogiids are adept at generating suction to feed (Bloodworth & Marshall, 2005), and it is known that kogiid hyolingual muscles are relatively large (Reidenberg & Laitman, 1994). However, the main finding of this study was that differences in theoretical maximum muscle tension, and vector analyses are probably insufficient to account for differences between kogiid and bottlenose dolphin feeding strategies. Although this study did show differences in hyoid surface area, and this was probably an important factor for generating suction in kogiids, the underlying functional mechanisms of increased hyoid surface area differ from that proposed by Heyning & Mead (1996).

The functional models from both genera demonstrate that the relative angular difference of hyoid and tongue depression–retraction between kogiids and bottlenose dolphins was only 6.3° and 1.5°, respectively. These differences are probably too minor to account for divergent feeding behaviours between the genera. The major difference between kogiids and bottlenose dolphins in vector analyses was the force of tongue retraction (58.8 and 174.4 N, respectively). Differences in vector values of the hyoglossus in bottlenose dolphins account for this functional difference. The greater MTMT of the hyoglossus in bottlenose dolphins may indicate a greater functional importance of the tongue to re-orientate food within the buccal cavity after prey capture than in kogiids. As bottlenose dolphins consume fish with sharp fin rays, re-orientation of prey is more important than in kogiids, which feed mainly on soft-bodied squid.

A combination of increased rapidity of hyolingual retraction, basihyal–thyrohyal surface area and specialized orofacial morphological features in kogiids were probably important for generating large negative intraoral pressures. The rapidity of hyolingual muscle activation is probably more important for generating suction than MTMT. Kinematic data demonstrate that kogiids depress and retract the hyoid more rapidly than bottlenose dolphins, which corresponds to greater suction capability (Bloodworth & Marshall, 2005). A study of muscle fibre type of hyolingual muscles could lend further support to rapid hyoid depression in kogiids and provide a further quantitative measurement of kogiid adaptation for rapid suction feeding. To maximize intraoral pressure development, the sternohyoideus, sternothyroideus and thyrohyoideus muscles probably depressed and retracted the hyoid rapidly, but also depressed the mandible via the geniohyoideus muscular insertion to the sternohyoideus. In addition, mandibular depression is likely to have been aided by a small digastricus muscle (von Schulte & Smith, 1918; Reidenberg & Laitman, 1994; Heyning & Mead, 1996). Hyoid rotation alone probably pulled the tongue caudoventrally, but simultaneous contraction of the hyoglossus and styloglossus muscles at, or slightly after, initial hyoid rotation most likely also increased the velocity of mandibular depression. The combination of rapid tongue depression and retraction, and rapid jaw opening resulted in a rapid increase in oral volume with a concomitant decrease in intraoral pressure. The hyoid's return probably resulted from contractions of the mylohyoideus, blubber elasticity and potentially contraction of the genioglossus. When the hyolingual apparatus forcefully returned to its resting state before gape closure, the effect was to jet water and food out of the mouth. This mechanism is supported by kinematic data in which hydraulic jetting events were frequently observed during feeding trials (Bloodworth & Marshall, 2005).

Although tongue morphologies of kogiids and bottlenose dolphins was not the focus of our study, it was noted that the tongues of kogiids were short rostrocaudally, but broad near their roots. This differed dramatically from the tongues of bottlenose dolphins, which were long and narrow. A broad, fused basihyal–thyrohyal complex probably facilitated the rapid depression and retraction of the entire kogiid tongue by providing a wider origin of attachment for lingual muscles. The rapid increase in intraoral volume was probably also assisted by the precise and tight fit of the kogiid tongue into a palatal vault. The increase in depth of the palatal vault would increase intraoral volume and the tight fit of the tongue in the vault could maximize volume change. A similar dependence on tongue and palatal vault morphology for generating suction has been documented in walruses, which are well known for their suction capability (Fay, 1982; Gordon, 1984; Kastelein et al. 1994). Our anatomical data correspond well with kinematic data for kogiids. During controlled kogiid feeding trials (Bloodworth & Marshall, 2005), the short, broad tongue acted as a piston, which was rapidly depressed and retracted during suction events.

In addition, our observations of odontocete tongues corresponded well with data from Werth (2006) regarding odontocete mandibles. The mandibles of kogiids are the bluntest and widest of all extant odontocetes (34 species; Werth, 2006), as are the skulls (Miller, 1923). The short, wide mandible and intermandibular space in kogiids (Werth, 2006) correspond to the short and wide tongues observed in this study. The blunt, wide, mandibles and tongues and increased breadth of the basihyal–thyrohyal elements of the hyoid apparatus are probably adaptations for rapidly displacing large intraoral volumes.

Kogiids possess other morphological adaptations that assisted in generating suction. In situ, the short mandible of kogiids was underslung and recessed within thick blubber (Fig. 7). The mandible is capable of opening nearly 90° (Bloodworth & Marshall, 2005). Due to its anatomy, the depressed mandible resembled a short trapdoor, distinctly different from bottlenose dolphins and other delphinids. In addition, the dense, fibrous, vertical ridges on the margins of the lower jaw effectively block lateral gape when the jaw was open, which resulted in a tube-like appearance when viewed from the front of the animal (Fig. 7). This surprising morphology is probably critical in maintaining negative pressures within the buccal cavity, and is analogous to the tube-like extension of the kinetic jaws of fishes that use suction extensively. Other marine mammals such as walruses (Fay, 1982; Kastelein et al. 1994, 1997b; Marshall et al. 1998) and bearded seals (C.D.M., unpublished data) are able to block lateral gape by pursing their lips and sealing the sides of their mouth using elaborated facial and perioral musculature. Because most cetaceans are not able to manipulate their lip margins, kogiids have evolved a novel, but effective solution to block lateral gape. Lastly, throat grooves in K. sima may permit additional gular distension, as hypothesized in other odontocete suction feeders (Clarke et al. 1968; Heyning & Mead, 1996; Werth, 2000a, 2004). The more pliable gular blubber of Kogia may allow greater depression of the hyolingual apparatus.

Fig. 7.

Still frames from underwater video of the lower jaw in kogiids. (A) Lateral view. The asterisk depicts the vertical ridge of tissue that occludes lateral gape by forming a tubular mouth. (B) Frontal view. The wide gape and short protruding mandible create a circular aperture.

The hyolingual complex and orofacial morphology of kogiids is probably just one of many solutions evolved by odontocete suction feeders. For example, despite the long, narrow rostra of beaked whales, these marine mammals are known to be suction feeders, and their hyolingual apparatus appears to be more robust than in delphinids (Heyning & Mead, 1996). However, they also possess interesting morphological adaptations of the blubber that border the lateral aspects of the mouth. Such structures are probably adaptations for reducing lateral gape and maintaining negative intraoral pressures along their long, narrow jaws. Although it appears that increased maximum theoretical muscle tension may not be important in kogiid suction feeding, orofacial and hyolingual morphologies (including enhanced tension potentials) in beaked whales may be fundamentally important to feeding in this poorly studied group, as hypothesized by Heyning & Mead (1996).

The hyolingual and oral anatomy of kogiids and bottlenose dolphins support the kinematic data (Bloodworth & Marshall, 2005), which verifies that kogiids use suction as their primary mode of feeding, whereas bottlenose dolphins and perhaps most delphinids are primarily ram feeders. Enlarged basihyals and thyrohyals are probably key in kogiid suction generation, and validate the functional hypothesis that enlarged hyoids are associated with suction feeding, but accomplish this by a mechanism other than differential muscle tension. These data serve to provide a foundation for future studies that can place odontocete feeding biomechanics within evolutionary and ecological perspectives.