SEARCH

SEARCH BY CITATION

Keywords:

  • CT scans;
  • cranial vasculature;
  • brain;
  • asymmetry

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. LITERATURE CITED
  8. Supporting Information

Evolution of endocranial anatomy in cetaceans is important from the perspective of echolocation ability, intelligence, social structure, and alternate pathways for circulation to the brain. Apart from the importance of studying brain shape and asymmetries as they relate to aspects of behavior and intelligence, cranial endocasts can show a close correspondence to the hydrostatic shape of the brain in life, and canals and grooves can preserve features of the circulatory system. Multiple samples are rarely available for studies of individual variation, especially in fossils, thus a first step in quantifying variation and making comparisons with fossils is made possible with CT scans of osteological specimens. This study presents a series of high-resolution X-ray CT-derived cranial endocasts of six extant species of Phocoenidae, a clade including some of the smallest and one of the rarest cetaceans. Degree of gyrification varies interspecifically and intraspecifically, possibly resulting from variation in preservation of the ossified meninges. Computed tomographic data show that visually assessed asymmetry in the cranial endocasts is not correlated with volumetric measurements, but nonetheless may reflect torsion in the skull's shape such that the right cerebral and cerebellar hemispheres extend rostrally and laterally more than the left. Vasculature and canals are similar to other described cetacean species, but the hypophyseal casts are unusual. Similarities between brain shape and volume measurements in the different species can be attributed to paedomorphism and concomitant variation in ecological preferences. This may explain similarities Neophocaena phocaenoides and Phocoena sinus share with the juvenile Phocoena phocoena specimen studied. Anat Rec, 296:979–992, 2013. © 2013 Wiley Periodicals, Inc.

Toothed whales (Cetacea: Odontoceti), particularly dolphins and their relatives (Odontoceti: Delphinoidea), are recognized as having high brain-to-body size ratios (encephalization quotient, or EQ, see: Edinger, 1955; Jerison, 1973; Marino et al., 2004a) and cortical reorganization and increased complexity relative to other mammals (Morgane et al., 1980; Marino et al., 2007). The cetacean brain has been compared to hominoid brains because of their similarly higher levels of encephalization, and complexity in gyrification, or superficial folding of gray matter leading to increased surface area of the brain, thought to be involved with increased intelligence (Edinger, 1955; Jerison, 1973; Marino et al., 2004a; Marino et al., 2007). Additionally, asymmetry of the brain in cetaceans has not previously been documented explicitly, possibly because of focus on whole brains removed from skulls (e.g., Morgane et al., 1980; Marino et al., 2003) that can cause distortion. In great apes (especially humans), brain asymmetry and torque is of considerable interest because it is correlated with behavioral traits such as handedness, language, and motor function (Toga and Thompson, 2003). In cetaceans, asymmetry of brain waves has been recorded as part of their unihemispheric sleep patterns (Ridgway, 2002). Furthermore, neuronal fiber composition and small size of the corpus callosum relative to brain size (Keogh and Ridgway, 2008) indicate the potential for asymmetrical use of each hemisphere. While much interest in the endocranium of cetaceans still exists, few studies have focused on gross morphological features in a monophyletic group of cetaceans, especially in the context of how it can aid studies of fossil specimens.

Secondary adaptations for survival in an aquatic existence, along with the evolution of high-frequency hearing and echolocation, are evident in the cranial anatomy of porpoises. Porpoises (Delphinoidea: Phocoenidae) are a clade of toothed whales (Odontoceti) comprising some of the smallest cetaceans, strongly supported as belonging to the Delphinoidea (e.g., McGowen et al., 2009), which includes oceanic dolphins (Delphinidae) and the beluga and narwhal (Monodontidae; see Fig. 1). Some controversy still exists regarding the relationships among porpoises, as well as their position among other delphinoids (e.g., Fajardo-Mellor et al., 2006a; Murakami et al., 2012b, 2012a), although many recent analyses find them sister to monodontids (McGowen et al., 2009; Steeman et al., 2009; Geisler et al., 2011). McGowen et al. (2009) estimate that phocoenids diverged from monodontids 15.5 million years ago, and the earliest known fossil phocoenid is ∼11 million years old (Barnes, 1985). Six (Rice, 1998) or seven (Committee on Taxonomy, 2012) extant phocoenid species are currently recognized: Neophocaena phocaenoides (and Neophocaena asiaeorientalis), Phocoena sinus, Phocoena spinipinnis, Phocoena dioptrica, Phocoena phocoena, and Phocoenoides dalli. The small size of porpoises is related to widespread paedomorphism in the group (Galatius, 2010; Galatius et al., 2011), which is particularly notable in the rarest species, P. sinus (Mellor et al., 2009). Researchers have proposed that because porpoises reach sexual maturity at earlier ages than other cetaceans, their recent evolution involved progenetic paedomorphosis (Ichishima and Kimura, 2005), or reducing developmental time to reproductive age at an immature stage relative to a nonpaedomorphic ancestor (Gould, 1977; Alberch et al., 1979). Galatius et al. (2011) proposed that paedomorphism in porpoises was promoted by reliable access to food sources, because small body size would confer high heat loss in the colder antitropical environments they inhabit, requiring the ability to acquire constant food. Phocoena dioptrica and Ph. dalli both prefer pelagic (open ocean) habitats (Barnes, 1985), while the remaining species prefer coastal, or in the case of the earliest diverging porpoise (N. phocaenoides), coastal and brackish or riverine habitats. Cranial differences related to progenetic paedomorphism are suggested to reflect the influences of habitat and constant access to food sources; for example, food is potentially less dependable in the pelagic habitats of the species (P. dioptrica and Ph. dalli) which have reverted to less paedomorphic morphologies (Galatius, 2010; Galatius et al., 2011). It is expected that endocranial morphology would have concurrently been affected.

image

Figure 1. Recent hypotheses for the relationships of Phocoenidae and other delphinoids, modified from molecular-based (McGowen et al., 2009) and morphological-based approaches (Fajardo-Mellor et al., 2006).

Download figure to PowerPoint

In spite of the importance of endocranial features from the perspective of brain evolution and function in this group, a paucity of data exists for cetacean brain external morphology, with most published material concentrating on the bottlenose dolphin, Tursiops truncatus (e.g., Morgane et al., 1980; Colbert et al., 2005; Oelschläger and Oelschläger, 2009), and with few detailed accounts of the gross morphology of the brain or of the endocranial cavity. Internal anatomy of the brain in select taxa has been well-studied recently using modern cross-sectional tools such as magnetic resonance imaging (MRI) and traditional histological techniques (e.g., Marino et al., 2003; Houser et al., 2004; Marino et al., 2004b; Oelschläger et al., 2008). The evolution of the delphinoid brain has not been studied or described in detail, except involving comparisons of volumetric changes throughout the evolution of odontocetes (Marino et al., 2004a). Cranial endocasts not only represent the shape of the brains, but also impart data pertaining to the vascular supply to the brain and other aspects of external morphology. The unusual patterns of cranial vasculature in cetaceans have been of interest from evolutionary and functional perspectives (McFarland et al., 1979; Geisler and Luo, 1998). Although the mode and distribution of changes in the circulatory anatomy across cetaceans is not well understood, in comparison to their artiodactyl relatives cetacean endocranial vasculature is marked by size reduction of major vessels such as the internal carotid artery, and presence of a hypotrophied endocranial rostral rete mirabile (Geisler and Luo, 1998). The rostral rete mirabile is a complex network of vessels extending from the thoracic rete mirabile (retia that develop along the spinal cord and within the vertebral canals), presumed to function as an alternative supply of oxygenated blood to the brain (McFarland et al., 1979; Costidis and Rommel, 2012). Geisler and Luo (1998) summarized the major changes to endocranial vasculature relative to the general terrestrial ungulate condition as follows: (1) acute reduction of the internal carotid artery; (2) loss of the stapedial artery and its branches; (3) absence of the capsuloparietal emissary vein; (4) an extensive rostral endocranial arterial rete (located caudoventrally, contained within the cavernous sinus, usually surrounding the hypophysis). A rostral and caudal pair of vessels originate from the rostral rete and directly supply the brain. They are homologous with parts of the circle of Willis; however, they supply the brain independently because they are not interlinked subdurally. Vasculature relevant to the current study involves the endocranial vasculature listed above, although extracranial and endocranial retia are complex and interconnected (Fraser and Purves, 1960; Geisler and Luo, 1998; Costidis and Rommel, 2012; Racicot and Berta, 2013).

Among the “penetrations” of the cranial cavity (i.e., canals containing both nerves and vascular structures), Colbert et al. (2005) observed reduced terminal nerves extending into the ethmoid region in Tursiops truncatus, which were previously examined in several taxa by Ridgway et al. (1987). Burrows and Smith (2005) observed similar projections corresponding to the ethmoid region in river dolphin endocasts (except Platanista gangetica) but interpreted them as more likely to be blood vessels. Recent work has focused on the presence and importance of olfactory bulbs and tracts in baleen whales (e.g., Thewissen et al., 2011), but they are otherwise absent among extant odontocetes (Morgane et al., 1980). The cranial hiatus is another feature of cetaceans associated with ventrolateral displacement of the ear bones (tympanoperiotics); it is defined by Mead and Fordyce (2009) as the confluent opening for the posterior lacerate foramen, jugular foramen, and internal acoustic meatus.

Meningeal ossification may impact the morphology of cranial endocasts, and aids in preservation of the hydrostatic shape of the brain and other tissues and vessels. Owen (1866) was the first to record the ossified meninges in a sperm whale (Physeter catadon), and noted their presence in a number of other species (e.g., delphinids, platypus). Nojima (1988) detailed the development of the ossified meninges from a spotted dolphin population (Stenella attenuata), finding that ossification of the tentorium cerebelli and falx cerebri begins at about 1 year of age, and that it is not likely to be a pathological condition given its prevalence among individuals and species. He notes that the potential function(s) of ossified falx cerebri and tentoria cerebelli remain unclear, but postulated a relationship with vitamin intake. Nojima (1988) observed that delphinid and phocoenid ossification of tentoria occurred postnatally, while ossified falx and tentoria are often present in neonatal mammals (cat, horse, and dugong) as well as in fetuses of Steller's sea lion and common seal. Colbert et al. (2005) discussed the utility of the ossified falx cerebri and tentorium cerebelli in documenting the presence and morphology of the straight sinus using CT scans. The dorsal meninges were interpreted as continuous (as lateral extensions) with the ossified falx cerebri and tentoria cerebelli. These dorsal meninges were inferred to have obscured the normally highly gyrencephalic brain morphology in delphinids (Colbert et al., 2005, see Fig. 2).

image

Figure 2. (A) Three-dimensional CT volume rendering in nonorthogonal orientation, cut posterior to the vertex, of Phocoena spinipinnis skull showing the ossified falx cerebri, partial ossified tentorium cerebelli, and part of the transverse sinus canal within the ossified meninges. (B) Isosurface of Phocoena phocoena skull showing locations of CT slices in (C) and (D). (C) CT slice showing dorsal ossified meninges. (D) CT slice showing ossified meninges and transverse sinus canal. Abbreviations listed in Table 2.

Download figure to PowerPoint

Because of their potential to offer insight into the evolution and development of sensory abilities, endocranial studies have recently become commonplace in a number of other taxa, most notably terrestrial mammals (Macrini et al., 2007a,b,c; Macrini, 2009; Rowe et al., 2011), birds (Ksepka et al., 2012), nonavian dinosaurs (Balanoff et al., 2010; Bever et al., 2011), and squamates (Olori, 2010). Access to the endocranial cavities of rare species such as Phocoena sinus (endemic to the northern portion of the Gulf of Mexico, and currently considered the rarest among extant cetaceans—see Rojas-Bracho et al., 2008), and juveniles provides a window into the development of endocranial features and their potential evolutionary and ecological implications. These data have thus far been difficult to obtain, largely because of the reluctance to use destructive techniques on rare specimens. Computed tomographic (CT) scans nondestructively allow digital examination of endocranial anatomy and comprise a permanent digital record that can be used by future researchers. For example, vascular supply to the brain in cetaceans can be used to study the evolution of alternate pathways of circulation to the brain, and becomes accessible with the use of CT scanning methods. In addition, the retia may be a source of phylogenetic information (Geisler and Luo, 1998), as well as providing important functional information useful for veterinary and military applications (Houser et al., 2010).

In this study we quantitatively and qualitatively examine the endocranial morphology of all extant porpoise species, including quantification of intraspecific variation using five adults and one juvenile specimen of the harbor porpoise, P. phocoena. Our results provide a framework from which to study the evolution of endocranial cavities in this and other cetacean groups.

MATERIALS AND METHODS

  1. Top of page
  2. ABSTRACT
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. LITERATURE CITED
  8. Supporting Information

Cranial endocasts of 11 phocoenid skulls were digitally extracted, using AmiraTM version 5, from high-resolution X-ray CT (HRXCT; described by Ketcham and Carlson, 2001) scans acquired at the University of Texas high-resolution X-ray CT Facility. Representatives of six extant phocoenid species (Rice, 1998) were sampled [Neophocaena phocaenoides (n = 1), Phocoena sinus (n = 1), Phocoena spinipinnis (n = 1), Phocoana dioptrica (n = 1), Phocoenoides dalli (n = 1), and Phocoena phocoena (n = 6) (although see Committe on Taxonomy, 2012)]. Five adult female skulls of Phocoena phocoena were selected based on similar external measurements and sex to study intraspecific variation, and one juvenile of unknown sex was selected for ontogenetic comparison. The following institutions loaned the specimens scanned in this study: California Academy of Sciences, San Francisco, California (CAS); Natural History Museum of Los Angeles County, Los Angeles, California (LACM); San Diego Natural History Museum, San Diego, California (SDSNH); University of California, Berkeley Museum of Vertebrate Zoology, Berkeley, California (MVZ). Table 1 identifies specimens by their institutional catalog numbers and details the scanning parameters employed. Terminology generally follows that of Mead and Fordyce (2009; Table 2). Comparisons were also made with a previously published endocast of T. truncatus (Colbert et al., 2005).

Table 1. Scanned specimens and associated locality and scanning parameters
SpeciesMuseum numberLocality informationInterslice spacing (mm)Interpixel spacing (mm)Total slices
Neophocaena phocoenoidesLACM 86041Japan, Honshu, Aichi Prefecture0.230.2246515
Phocoena dioptricaLACM 86042Argentina, Tierra del Fuego Nat'l Terr0.330.2832441
Phocoena phocoenaMVZ 135247CA, Monterey Co.0.50.2734464
Phocoena phocoenaMVZ 90696CA, Contra Costa Co.0.720.1758423
Phocoena phocoenaMVZ 172161CA, Sonoma Co.0.720.1758423
Phocoena phocoenaMVZ 172160CA, Sonoma Co.0.720.1758402
Phocoena phocoenaMVZ 208560CA, San Mateo Co.0.720.1758402
Phocoena phocoenaMVZ 172141CA, Marin Co.0.220.1953419
Phocoena sinusSDSNH 21555Gulf of California0.210.1924493
Phocoena spinipinnisLACM 72442Peru, Puntas Chancoy0.330.2832373
Phocoenoides dalliCAS 15278CA, San Mateo Co.0.50.2686568
Table 2. List of abbreviations and definitions
AbbreviationElement
Bbasisphenoid
BObasioccipital
ccscarotid canal of the sphenoid
cgchiasmatic groove
chcranial hiatus
dsssdorsal sagittal sinus sulcus
dsscdorsal sagittal sinus canal
egvcnendocranial groove for the vestibulocochlear nerve
EOexoccipital
Ffrontal bone
fg/ocfrontal groove cast/optic canal cast (includes inferior & superior orbital fissures)
fmforamen magnum
frforamen rotundum
hchypoglossal canal
hfhypophyseal fossa
iorminternal opthalmic rete mirabile cast
isintersphenoidal synchondrosis (or groove on endocast corresponding to raised presphenoid-basisphenoid articulation)
MEmesethmoid
mmamiddle meningeal arteries
Nnasal bone
nonasal openings (cross section)
ofcossified falx cerebri
otcossified tentorium cerebelli
Ppresphenoid
ptipyramidal tract impression (cast)
scsinus confluens
smaspinal meningeal artery traces
sscsagittal sinus canal
tsctransverse sinus canal
Vvomer

Cranial endocasts were digitally extracted (i.e., segmented) using the “magic wand” and “paintbrush” tools in AmiraTM. Cavity openings were bounded curved planes defined by the edges of the openings. Original CT data and STL files of the endocasts are available upon request from the authors.

To compare interspecific and intraspecific proportional differences among the cranial endocasts, measurements were standardized to cranial volume or corresponding linear measurements of the skull (Tables 3 and 4). “Landmark” and “measure” tools were used to obtain standard cranial and endocranial measurements in AmiraTM and AvizoTM (i.e., greatest skull and endocast lengths, greatest skull and endocast widths, and greatest skull height, see Table 4). To measure from the midline of the cranial endocast laterally, a landmark was placed at the separation of the two cerebral hemispheres. For cerebellar and other hindbrain measurements, a landmark was placed at the center of the pons cast. The volumes of the crania (after Mead and Fordyce, 2009), excluding the mandibles and tympanoperiotics, and endocasts, were obtained in VGStudioMaxTM 2.2 using the “Object info” under “Properties” of each scan, after performing “surface determination.” The “polyline 3D” tool was used to obtain the cranial volume caudal to the antorbital notches (Supporting Information Fig. 1). Because we did not have individual body size/mass measurements for all specimens, cranial volumes were employed as the only available option for standardizing endocasts for comparison. The volume of the cranium, exclusive of cavities and the rostrum, as opposed to entire skull volumes, was used to avoid introduction of additional voxels from varying numbers of teeth in each specimen (although the differences in results between using skull and cranium to standardize endocast volumes were minimal, see Table 3). To determine whether qualitatively observed asymmetry was reflected in volumetric measurements, linear and volumetric measurements were taken of the two cerebral and cerebellar hemispheres in both AvizoTM and VGStudioMaxTM. In VGStudioMaxTM, the 3D polyline tool was used to separate the left and right hemispheres of the whole endocast (including the brainstem and cerebellum). Plots comparing endocast volumes were made with the statistical software R (R Development Core Team, 2012). The standardized volumes of the endocranial cavities were delineated into characters and parsimony ancestral state reconstructions on a morphology-based tree (Fajardo-Mellor et al., 2006b) and molecular-based tree (McGowen et al., 2009) were performed in Mesquite v. 2.75 (Maddison and Maddison, 2011) with characters unordered. Character states were polarized using the outgroup comparison method (Watrous and Wheeler, 1981), with T. truncatus as the outgroup, because currently it is the only specimen for which we have comparable data.

Table 3. Volume measurements of phocoenid skulls, crania, and cranial endocasts
SpecimenLeft volume (cc)Right volume (cc)Endocast total volume (cc)Skull volume (cc)Cranial volume (excludes rostrum) (cc)Endocast/skull volumeEndocast/cranial volume
P. phocoena MVZ135247278.99284.3563.29545.15434.031.031.30
P. phocoena MVZ172161254.89255.7510.59478.384387.9241.071.32
P. phocoena MVZ208560277.972287.755565.727420.23338.771.351.67
P. phocoena MVZ90696285.59272557.59448.38353.841.241.58
P. phocoena MVZ172160260264.53524.53441.4350.791.191.50
N. phocaenoides259.718251.623511.341270.083224.791.892.27
P. sinus180.14177.32357.46173.72147.052.062.43
P.spinipinnis271.47287.9559.37448.75368.411.251.52
P.dioptrica239.95242.61482.56507.413443.4030.951.09
P.dalli483.377467.011950.388668.59551.861.421.72
P. phocoena MVZ172141 (juv)144.67139.51284.18116.4496.542.442.94
T. truncatus (Colbert et al., 2005)not measurednot measured2131.883098.5682292.960.690.93
Table 4. Linear measurements of cranial endocasts of all specimens
SpecimenEndocast total length (mm)Endocast total width (mm)Endocast total height (mm)Right cerebrum width (mm)Left cerebrum width (mm)Right rostrocaudal cerebrum length (mm)Left rostrocaudal cerebrum length (mm)Right dorsoventral endocast height (mm)Left dorsoventral endocast height (mm)Right cerebellar width (mm)Left cerebellar width (mm)Medioalateral width presphenoid-basisphenoid groove (mm)Mediolateral width pds/endocast width
P. phocoena MVZ13524790.79132.1298.5266.4566.8289.891.2446.4947.631.1428.8118.090.14
P. phocoena MVZ17216188.86125.5292.6755.8963.6389.7988.8160.4757.792828.6910.80.09
P. phocoena MVZ9069691.67128.2994.8867.4567.5692.891.1157.5659.8430.7530.5415.010.12
P. phocoena MVZ9069691.72130.2395.4466.0567.190.3788.2845.6748.730.7930.7917.90.14
P. phocoena MVZ17216088.99138.4595.2762.5464.685.1486.950.4352.7731.8832.169.470.07
N. phocaenoides98.80118.7298.8361.560.0792.9293.5362.361.8829.4828.58.610.07
P. sinus78.47106.1085.5159.3960.7472.973.4156.1355.3224.8524.429.420.09
P. spinipinnis94.78128.2196.6563.2762.491.5988.9950.2851.2227.8328.885.790.05
P. dioptrica94.78128.2196.6563.2762.491.5988.9950.2851.2227.8328.885.790.05
P. dalli113.49161.43105.3385.1784.97107.9108.7867.1369.8845.3444.1919.010.12
P. phocoena MVZ172141(juv)74.9999.6474.6456.255.7671.8872.7435.12364819.2420.6612.660.13

RESULTS

  1. Top of page
  2. ABSTRACT
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. LITERATURE CITED
  8. Supporting Information

General Descriptions of Endocasts

The presence of ossified tentoria cerebelli and falx cerebri (Fig. 2) allowed digital representation of not only the cranial endocasts but also casts representing aspects of the cranial vasculature. The reduced carotid canals of the sphenoid typical of odontocetes are visible in all specimens (Fig. 3, ccs), as well as the dorsal sagittal sinus canal, dorsal sagittal sinus sulcus, sinus confluens, and transverse sinus canals. The intracranial extensions of the spinal meningeal arteries traverse the dorsolateral surface of the cerebellum and rostral portions converge to become the internal ophthalmic rete mirabile, visible on the ventral surface of the endocranial cavity rostral to the pontine and pyramidal tract impressions (Fig. 3, iorm, sma). The middle meningeal arteries are thin and extend dorsally along the Sylvian sulcus (Fig. 3, mma). Penetrations, sensu Colbert et al. (2005), which in life would have conducted neural and circulatory tissue were also extracted (Fig. 3, casts highlighted in blue). The rostral-projecting penetrations consist of the frontal groove/optic canal casts, cast of the foramen rotundum and maxillary branches of the trigeminal nerve canals. The endocranial groove for the vestibulocochlear nerve is visible on the surface of the cerebellum communicating rostrally with the cranial hiatus, an undifferentiated space that houses the tympanoperiotic bones, peribullary sinus, and that also represents a confluence of the jugular foramen, posterior lacerate foramen, internal acoustic meatus, and sometimes the foramen ovale (Mead and Fordyce, 2009). The hypoglossal canals are reduced, consistent with other odontocetes, and positioned caudal to the cranial hiatus (Fig. 3, hc).

image

Figure 3. (A) Digitally isolated cranial endocast of a harbor porpoise (Phocoena phocoena) in the context of the skull at 45° angle; skull rendered transparent in dorsal (i) and left lateral (ii) views. (B) Orthogonal views (from top left to right: anterior, dorsal, left lateral views; from bottom left to right: posterior, ventral, and right lateral views) of the same digital cranial endocast showing “penetrations” (sensu Colbert et al. 2005, includes canals for nerves) segment in blue, circulatory segment in red, and neural portion in pink. Abbreviations listed in Table 2.

Download figure to PowerPoint

To compare with the delphinid T. truncatus (Colbert et al. 2005), we extracted the volumes of the circulatory and penetrations portion in one individual of P. phocoena (MVZ 135247). The volume of the neural portion of P. phocoena specimen MVZ 135247 is 552.98 cc, making up 82.98% of the total endocast volume, while the circulatory segment (6.37 cc) is 0.96% of the endocast volume, making it the smallest proportion. The penetrations portion comprises 107.04 cc, or 16.06% of the total endocast volume.

The juvenile's total endocast volume (284.19 cc) is closest to that of P. sinus (357.15 cc, Fig. 4, Table 3). Phocoenoides dalli has the largest endocast volume (950.04 cc), while P. phocoena, P. spinipinnis, and P. dioptrica cluster together with similar volumes. If the volume of bone in the cranium is a reliable proxy for overall volume of the head region in porpoises, then the ratio of endocranial volume to skull volume can be taken to represent standardized brain size among specimens. Ordinary least squares regression of endocranial volume and skull volume on unlogged axes reveals reasonably strong (and statistically significant) linear correlation (R2 = 0.5332, P < 0.01); including juvenile measurements increases both the strength and significance of this correlation (R2 = 0.6539, P < 0.01; Fig. 4), indicating some potential proportional differences in relative brain volume. Given this, relative brain size (endocast volume standardized to cranium volume) is consistent among adult P. phocoena (Fig. 5); however, relative brain size in the juvenile is disproportionately large, and most closely resembles P. sinus, and to a lesser degree, N. phocaenoides. Phocoena spinipinnis, Ph. dalli, and P. dioptrica fall approximately within the same range as adult P. phocoena specimens.

image

Figure 4. Bivariate plot of endocast volume versus cranial volume. Regression lines including and excluding juvenile almost completely overlap.

Download figure to PowerPoint

image

Figure 5. Bar graph comparing endocast volume standardized by cranial volume of all specimens studied.

Download figure to PowerPoint

The dorsal surface of the juvenile's endocast is mostly smooth, while all of the adult specimens have gyrencephalization that varies in amount of visible folding (Fig. 6). Phocoena phocoena specimens MVZ 208560, MVZ 172160, and MVZ 90696, and P. dioptrica, P. spinipinnis, and N. phocaenoides all reflect extensive impressions of sulci and gyri, especially visible dorsally in the rostral region of the telencephalon (Figs. 6 and 7). Ventrally and rostrally in the temporal and frontal regions all specimens show similarly dense gyrification, but P. phocoena specimens MVZ 172161 and MVZ 135247, and Ph. dalli and P. sinus have less obvious indications of sulci and gyri, especially caudally. The CT slice data for P. sinus show that it has dorsal ossified cerebral meninges, however others specimens do not exhibit them as extensively (i.e., juvenile P. phocoena, N. phocaenoides, Ph. dalli). All specimens have grooves indicating the prominent Sylvian (lateral) sulci, although in the juvenile it is more shallowly delineated than in adults.

image

Figure 6. Standard orthogonal views (views from left to right: anterior, posterior, dorsal, ventral, left lateral and right lateral) of described Phocoena phocoena specimens (museum number indicated). Scale applies to all specimens.

Download figure to PowerPoint

image

Figure 7. Standard orthogonal views (views from left to right: anterior, posterior, dorsal, ventral, left lateral, and right lateral) of all described phocoenid species examined. Scale varies among specimens.

Download figure to PowerPoint

Visually, some asymmetry is observed among all specimens. The left temporal region is more laterally and rostrally expanded than the right, although the linear and volumetric measurements of this region are subequal (Figs. 6 and 7, Tables 3 and 4). The left temporal region is also more obviously delineated from the rest of the cerebrum in dorsal view. The right cerebral hemisphere appears to extend dorsally more than the left, again not reflected in volumetric and linear measurements (Figs. 6 and 7, Tables 3 and 4). The juvenile and adult P. phocoena, and Ph. dalli show similar patterns to the other species but appear more symmetrical overall (Figs. 6 and 7). Neophocaena phocaenoides, P. sinus, P. spinipinnis, and P. dioptrica all have laterally but not caudally expanded left cerebral hemispheres and more dorsally and caudally expanded right cerebral hemispheres. Neophocaena phocaenoides has the most transversely compressed cerebrum of all the species. The cerebellar region in all species/individuals is moderately symmetrical by comparison. We examined the strength of correlation between left and right endocast volumes using both a rank-based nonparametric method (Spearman's rho), and ordinary least squares regression (Fig. 8). The results of nonparametric correlation tests reveal strong and statistically significant correlation (Sp. rho = 0.9855; P < 0.001); regression of these data further reveals a significant linear relationship (R2 = 0.9, P < 0.001). These results indicate low volumetric asymmetry between left and right sides, and in fact data fall extremely close to an idealized 1:1 line (dashed line in Fig. 8).

image

Figure 8. Bivariate plot showing left versus right halves of endocast volumes. Solid line represents slope of left versus right volumes, dashed line represents slope of 1:1 correlation.

Download figure to PowerPoint

Cerebral Casts

A rostromedial ridge projecting dorsoventrally along the mesethmoid and presphenoid (Fig. 9, thick arrow) defines a groove separating the left and right cerebral hemispheres in all specimens, visible in anterior views at the rostral end of the endocasts (Figs. 6 and 7). It is shallowest in N. phocaenoides (where it is most distinct dorsally) and P. spinipinnis and the juvenile P. phocoena.

image

Figure 9. (A) Three-dimensional reconstructed volume of Phocoenoides dalli showing ventral endocranial cavity with dorsal skull removed, as shown in (a). Thick arrow points to the rostromedial ridge lying along the mesethmoid and presphenoid; (B) Sagittal CT-scan slice of Phocoena phocoena with labeled endocranial anatomy, position of slice shown in (b). Abbreviations listed in Table 2.

Download figure to PowerPoint

At the intersphenoidal synchondrosis, the presphenoid and basisphenoid are raised, (Fig. 9, p-bs) corresponding to distinct rostromedial grooves on the ventral cerebrum casts (Fig. 3, p-bs). The grooves are visible in most specimens to varying degrees (Figs. 6 and 7). The intersphenoidal synchondrosis grooves of P. dioptrica, Ph. dalli, and P. phocoena (except MVZ 172160, with a groove at 9.47-mm long, ∼7% endocast width) are transversely elongate rectangular grooves (Table 4), while in P. sinus the groove is more rounded at its ends. In N. phocaenoides and P. spinipinnis it is relatively transversely narrow (8.6 mm, 7% endocast width; 5.7 mm, 4.5% endocast width, respectively—remaining species are between 8 and 12.7% endocast width, see Table 4). Posterior to this groove in all specimens is the shallow optic chiasm/chiasmatic groove (Figs. 6, 7, and 9). The chiasmatic groove casts (manifest as eminences on the endocasts) are continuous with the optic canals, which are quite distinct in P. dioptrica, Ph. dalli, and P. spinipinnis. In Ph. dalli the optic tract comprises a trapezoidal platform distinct from other endocast regions, the posterior borders of which contain small casts of the carotid canals of the sphenoid. The optic tracts in N. phocaenoides, P. sinus, and P. phocoena are less differentiated.

Rounded eminences representing the hypophyseal fossae posteromedial to the carotid canal casts are most obvious in P. dioptrica, and only slightly less prominent in P. spinipinnis (Fig. 7). This eminence is a shallower peanut-shaped cast in Ph. dalli. In the remaining specimens, the casts are less prominent but still transversely wide. Caudal to the elliptical-shaped, shallow, hypophysis cast in N. phocaenoides is a small but distinct notch representing the dorsum sellae and the sphenoccipital synchondrosis. In all specimens a similar notch represents the termination of the hypophyseal fossa.

Marked differences are present between the adult and juvenile P. phocoena cerebral endocasts. The juvenile's temporal lobes do not expand laterally as much as in the adults, but the frontal regions are developed laterally almost to the full extent seen in adults (Table 4). Phocoenoides dalli and P. dioptrica have more sharply inflected rostral aspects to their cerebra than the other taxa. Both Ph. dalli and P. dioptrica appear to have the most rostrocaudally compressed endocasts in dorsal view (Fig. 7), but the rostral extent of the P. dioptrica endocast is more rounded in dorsal view. Phocoena dioptrica is the only specimen whose anterior cerebrum has a rounded rostral tip, the remainder are more squared in appearance rostrally.

Hindbrain Casts

The cerebellar casts are proportionally smaller in the juvenile than in the adults, and are not particularly prominent compared with the cerebral casts (Table 4). Their size and position in the juvenile compared with other specimens is most similar to N. phocaenoides and P. sinus. While smaller than the cerebrum in the adults, the cerebellar casts still make up a large proportion of the neural segment, and are distinctly delineated from the cerebrum by the ossified tentoria cerebelli. The pyramidal tract is visible to varying degrees among adult individuals of P. phocoena, as well as in P. spinipinnis, and Ph. dalli, with bulges representing distinct pons and medulla oblongata casts standing out the most among these individuals. The juvenile P. phocoena, N. phocaenoides, P. sinus, and P. dioptrica do not have prominent pons and pyramidal tract casts. Neophocaena phocaenoides has the least differentiated ventral aspect of the hindbrain.

Circulatory Casts

The spinal meningeal artery traces are visible and elaborated on the juvenile specimen, and presence of the dorsal sagittal sinus and the sinus confluens is observed, although neither is as prominent or well defined as in the adults of all species. The rostrodorsal-extending wisps visible in the juvenile specimen are traces of the ethmofrontal suture (Fig. 6). All sampled specimens exhibit a sinus confluens except P. phocoena (MVZ 135247), which has a shunt between the two transverse sinuses. Phocoena spinipinnis has the most elaborated lateral extents of the spinal meningeal arteries, with two to three interconnecting branches, while the remaining specimens have one large vessel with smaller branches that subsume rostroventrally, connecting with the highly reduced carotid canals of the sphenoid. The dorsal sagittal sinus casts lie either on the right or left cerebral hemisphere casts along their dorsal traverse between the interhemispheric fissure, with no obvious trend among all the specimens. In N. phocaenoides, P. spinipinnis, P. phocoena juvenile, MVZ 172161 the dorsal sagittal sinus cast lies nearer to the right hemisphere. In P. phocoena MVZ 135247, MVZ 208560, MVZ 90696, MVZ 172160, Ph. dalli, P. dioptrica, and P. sinus it lies nearer to the left. The dorsal sagittal sinus cast terminates in the mediolaterally thin, shallow dorsal sagittal sinus sulcus cast. The middle meningeal artery casts are rostrocaudally thin and visible to varying degrees in all the specimens, with more obvious impressions in the juvenile reaching to the dorsal sagittal sinus cast.

“Penetrations” Casts

The P. phocoena juvenile's hypoglossal canal casts extend from a more caudal position of the cerebellar region than in any of the adults. Their position is most similar to those of N. phocaenoides (Figs. 6 and 7). The frontal groove/optic canal casts are not as apparent or developed in the juvenile P. phocoena specimen. The frontal groove/optic canal casts are rostrocaudal elliptical canals most obvious in MVZ 135247 and MVZ 172161. They are similar in shape in P. dioptrica, but not as well formed in the other species. Phocoena dioptrica and Ph. dalli have many well-developed canals in the penetrations portion of their cranial endocasts, whereas the other species do not. Cranial nerve “0” is visible in the CT scans of the specimens, but was not specifically digitally extracted.

Character Optimization

Optimization of relative endocast volumes on both phylogenies (Fig. 10) indicates that the most parsimonious pattern of evolution involves three transitions, with the relatively largest endocast volumes evolving independently in N. phocaenoides and P. sinus. The condition of the hypothetical ancestor of the phocoenid lineage is equivocal, possessing any of the three standardized volumes in either morphological or molecular topologies. In both topologies, the hypothetical ancestor of phocoenids excluding N. phocaenoides has an endocast volume at least 75% of the skull volume. In the molecular phylogeny, a clade including P. sinus, P. spinipinnis, and P. dioptrica share the inferred state of the ancestor of all phocoenids excluding N. phocaenoides. In both phylogenies P. phocoena and Ph. dalli's common ancestor has an endocast volume between 75% and 150% of the skull volume; in the morphological phylogeny P. dioptrica's common ancestor with these two species also shares that character state.

image

Figure 10. Parsimony ancestral state reconstructions of endocranial volumes on morphology-based tree (Fajardo et al., 2006) and molecular-based tree (McGowen et al., 2009). Characters unordered. Character state 0 = standardized volume less than 1; character state 1 = standardized volume 1–2; character state 2 = standardized volume greater than 2.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. ABSTRACT
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. LITERATURE CITED
  8. Supporting Information

The volumetric and linear measurements contradict obvious visual/morphological differences between the two hemispheres in the specimens studied. Accordingly, the asymmetry among specimens may reflect torsion, as has been described in archaeocetes (Fahlke et al., 2011). Torsion of the skull leading to directional asymmetry may have been a critical adaptation in cetaceans relating to the development of biosonar (Fahlke et al., 2011), or alternatively may be a by-product of changes in arrangement of the larynx to allow swallowing underwater (Macleod et al., 2007). The discovery of cerebral and cerebellar asymmetries among phococenid endocasts reinforces the potential for behavioral similarities observed in great apes and other animals (Toga and Thompson, 2003), as well as previously observed asymmetrical blood flow in other odontocetes (Houser et al., 2010). The torsion observed here may also be correlated with cerebral independence allowing unihemispheric sleep (Ridgway et al., 2006, 2009) and asymmetrical beam formation for echolocation (although they are symmetrical soft tissue structures, the right phonic lips are preferentially used to emit beams for orientation and foraging) and possibly communication (Madsen et al., 2010; Koblitz et al., 2012). Extant phocoenids are traditionally viewed as showing little external asymmetry compared to other odontocetes (Barnes, 1985; Ichishima and Kimura, 2005), in line with what we observe in linear and volumetric measurements. The asymmetry observed visually in this study is superficially correlated with shape of the pterygoid sinuses (Racicot and Berta, 2013), hinting at the complex relationship during ossification with soft-tissues. Significant volume differences between the hemispheres is not necessarily expected (Toga and Thompson, 2003), and perhaps analyses recently developed for primate skulls would perform better in documenting the asymmetry in endocasts or skulls of other taxa (Balzeau and Gilissen, 2010). Future studies could include detailed measurements of the degree of asymmetry in phocoenid and other odontocete endocasts, including those of fossil taxa, and how it may be reflected in other aspects of the skull.

Many of the specimens in this study exhibit notable gyrification, in contrast to a similar study on T. truncatus where the partially ossified dorsal meninges cover the impressions of sulci and gyri (Colbert et al., 2005). The degree of preserved gyrification may be related to variation in preparation of the dry skulls or individual variation in retention of ossified meninges. Some individuals may never exhibit partially ossified meninges such as found in T. truncatus (Colbert et al., 2005) relating to their paedomorphic tendency to generally reduce ossification (Barnes, 1985; Galatius, 2010; Galatius et al., 2011). This explanation is contradicted by the observation that P. sinus, otherwise the most paedomorphic porpoise (Mellor et al., 2009), has well-developed meninges (including the ossified falx cerebri and tentorium cerebelli) that obscure caudal gyrification. Alternatively, Marino et al. (2003) described a P. phocoena brain as having less elaborated dorsal and caudal areas of the brain, which may be correlated with reduced gyrification. Similarly, the diastema between the occipital lobes proposed by Marino et al. (2003) to be associated with reduced proliferation of tissue is also notable among all of our specimens except N. phocaenoides, which looks most like an endocast of T. truncatus (Colbert et al. 2005). Nonetheless, our analysis shows that the extent of gyrification does not represent a phylogenetically or ontogenetically informative character in phocoenids. Extensive gyrification indicates increased intelligence by increasing the surface area of grey matter in the brain (Prothero and Sundsten, 1984), but if meninges obscure this region in some individuals then we cannot use endocast gyrification to assess intelligence across taxa. Our study, however, provides a baseline for means by which to study the patterns of gyrification among the rare phocoenids, fossils, and other related taxa.

The relative size of the cerebellum to the cerebrum is smaller in the juvenile P. phocoena than in adults, similar to that observable in infant T. truncatus (Marino et al., 2004b). The juvenile's frontal lobe region does not extend rostrally as much as the adults (similar to an adult T. truncatus, Colbert et al., 2005). The volume of the juvenile's endocast (322.95 cc) is more than half of the published volume derived from MRI scans of an infant (estimated 2- to 3-months old) T. truncatus brain (705.3 cc, Marino et al., 2004b). The estimated volume based on an MRI scan of P. phocoena (Marino et al., 2003) falls within the lower end of our range of endocast volumes of adult P. phocoena individuals. Our larger volumes compared to these studies are likely related to inclusion of the entire endocranial cavity and associated structures, as in Colbert et al. (2005), and have the potential to be a more accurate reflection of this volume because the scans are higher-resolution. Compared with the volumes reported in Colbert et al. (2005), P. phocoena has a larger “penetrations” portion of the cranial endocast, and much smaller circulatory portions. Standardized volumes of phocoenid endcoasts are much greater than those of T. truncatus in Colbert et al. (2005), possibly relating to phocoenid paedomorphism (paedomorphism has been proposed as a mechanism by which anthropoid primates acquired higher encephalization quotients—see Vrba, 1996). Inclusion of representatives of hypothesized sister taxa (Monodontidae), as well as fossil species, may help resolve some of these questions.

The cast of the hypophyseal fossa is a prominent, seemingly transversely paired feature in the specimens of this study where it was most obvious, and is otherwise transversely elongate, a condition described previously in delphinid hypophyses (Cowan et al., 2008). Cowan et al. (2008) found dolphin hypophyses to be at least twice as transversely wide as they were rostrocaudally long, and the sella turcica to vary in distinctiveness among individuals. In an endocast of T. truncatus (Colbert et al., 2005), it is a shallow sphere, and in McFarland et al. (1979) the hypophysis is transversely elongated and rather similar to our results. Costidis and Rommel (2012) described the cavernous sinus as surrounding the sella turcica and hypophysis in T. truncatus as in many terrestrial domestic mammals, thus some phocoenids may exhibit greater endocranial distinction of the cavernous sinuses than previously documented in T. truncatus. Individual variation may also explain why the hypophysis appears superficially paired in some of the phocoenid specimens in this study. Dissections and whole head scans would be an appropriate approach to further investigate these findings, as limited literature exists on the topic, and it is of interest from the perspective of the abilities for marine mammals to maintain body temperature in relatively low temperature aquatic environments. The mid-ventral surface of the cerebral portion of the endocast in phocoenids appears to be less invested by retia than in T. truncatus (Colbert et al., 2005), which may be why the hypophyseal fossa and optic tracts appear more differentiated. Prominent olfactory lobes had been reported previously in P. phocoena by Morgane et al. (1980), which may restrict the development of the retia in this region.

As Burrows and Smith (2005) found with river dolphins, the dorsal sagittal sinus is asymmetrically situated but does not preferentially bend to the right or left among studied specimens. The shunt at the sinus confluens observed in P. phocoena may allow alternate drainage to each transverse sinus from the dorsal sagittal sinus; however it was only observed in one specimen among the individuals studied, and is thus a variable structure. The falx cerebri and tentorium cerebelli in all of the specimens were not as profoundly ossified as in T. truncatus (Colbert et al., 2005), so the straight sinus was not as well-defined in these scans, making it difficult to determine into which transverse sinus each vessel drained.

Neophocaena phocaenoides, P. sinus, and the juvenile P. phocoena show similarities in relative endocast volume and hindbrain size. Either the presumed adult N. phocaenoides and P. sinus are not fully mature individuals, or they show more pronounced paedomorphic traits than other phocoenid species. Morphological similarities between N. phocaenoides and P. sinus were also recovered by Galatius et al. (2011), which they relate to their relative near-shore or shallow water preferences, and consequent higher degree of paedomorphism resulting from access to abundant food sources. Character optimizations did not resolve this issue, suggesting that P. sinus and N. phocaenoides independently derived greater relative endocast volumes. Under the phylogenetic interpretations, the ontogenetic development of P. phocoena's standardized endocast volume potentially tracks the evolutionary trajectory of phocoenids toward slightly smaller relative endocast volumes. Because P. sinus are inbred and suffered a historical population bottleneck and have always had a low population size (Taylor and Rojas-Bracho, 1999; Rojas-Bracho et al., 2006), there may be an adaptive aspect to progenetic paedomorphosis or to the emphasis of existing paedomorphism.

On the other hand, Ph. dalli and P. dioptrica endocasts were expected to have similar morphologies given geometric morphometric similarities shown by Galatius et al. (2011), who suggested that their pelagic habitat preferences led to larger body size and extended maturation time compared to other phocoenids (i.e., they became less paedomorphic). For example, the compressed rostral aspects of the endocasts of Ph. dalli and P. dioptrica are likely correlated with the more posterior positioning of the skull vertex in these two species, as found by Galatius et al. (2011). The repositioning of the foramen magnum to a more dorsal position (Galatius et al., 2011) may also influence the brain's shape, particularly the cerebellum. As suggested by Ichishima and Kimura (2005), it is likely that the environment during early phocoenid evolution was fluctuating, and that progenetic paedomorphosis was an adaptive strategy. Consideration of fossil species may influence interpretations and help tease out the complexities of the evolution of paedomorphism in this group.

In summary, we performed an inter- and intra-specific qualitative and quantitative analysis of the cranial endocasts of the six extant porpoise species. We find that the proportions of circulatory and “penetrations” segments are quite small compared to the brain volume, which has potential implications for volumetric measurements and comparisons between brain-only studies and those using endocasts, although the circulatory and penetration portions are most certainly underestimates of actual cerebrovascular volumes. In addition, we find that many porpoise species exhibit gyrification on the endocranial surface. Although degree of gyrification can indicate increased intelligence, the intraspecific variation within the five examined adult specimens of P. phocoena is large, necessitating further study to establish whether gyrification is phylogenetically informative. We find that relative brain size (estimated as a proportion of total skull volume) differs among the studied species, and appears to differ significantly between juvenile (proportionately large) and adult (proportionately small) P. phocoena. Other species with proportionately large endocranial cavities are P. sinus and N. phocaenoides, supporting results from other authors that these species may exhibit extreme paedomorphic characters relating to their affinity for more shallow-water/coastal habitats. Future work on endocranial cavities in cetaceans should focus on broader-scale phylogenetic comparisons including measurement and gross anatomical description to further test the relationships between skull shape, endocast shape, and functional and ecological factors that can affect the evolution of endocranial space.

ACKNOWLEDGEMENTS

  1. Top of page
  2. ABSTRACT
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. LITERATURE CITED
  8. Supporting Information

The authors thank museums and staff for access to specimens and loaning specimens for scanning (D. Janiger and J. Dines of LACM; P. Unitt of SDSNH; J. Mead and C. Potter of USNM; C. Conroy of MVZ; and M. Flannery of CAS). Simon Darroch is thanked for assistance with R software and reviewing earlier versions of this work. Allison Hsiang is thanked for advice on format of some figures. Brian Beatty is thanked for discussion on ossified meninges. Tursiops truncatus scans were generously provided by digimorph.org. Constructive comments from the editor, J. Laitman, and two anonymous reviewers greatly improved the quality of the final manuscript.

LITERATURE CITED

  1. Top of page
  2. ABSTRACT
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. LITERATURE CITED
  8. Supporting Information

Supporting Information

  1. Top of page
  2. ABSTRACT
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. LITERATURE CITED
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
ar22704-sup-0001-suppFig1.tif14245KSupplementary Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.