CRANIAL, ORBITAL RETE, AND RED MUSCLE VEIN ANATOMY
INDICATE A HIGH DEGREE OF ENDOTHERMY IN THE BRAIN
AND EYE OF THE SALMON SHARK, LAMNA DITROPIS
Victor Tubbesing
Advisor: Barbara Block. Ph.D.
Hopkins Marine Station Spring Project
June 6, 1997
Permission is granted to Stanford University to use the citation and abstract of this paper.
ABSTRACT
Geographic range, extent of red muscle, supra-hepatic rete size, and
limited temperature measurements have placed the salmon shark, Lamna
ditropis, as the warmest of the lamnid sharks. Yet its anatomy has remained
largely undescribed, and measurements of brain or eye temperatures in
particular have not been taken. In this study, three specimens were examined
to determine if the morphological requirements for a warm brain and eyes are
present. In addition to local heat-producing red muscle of the eye and jaw,
other notable features were found. A well-developed orbital rete lies within a
venous sinus on both sides of the cranium. Cool, oxygenated blood from the
gills can pass through the dividing vessels of this exchanger before reaching the
brain or eyes. Since venous blood in the sinus flows opposite the arterial
blood, counter-current heat exchange can occur. A vein originating in the red
swimming muscle potentially contributes to the warmth of the venous sinus
by supplying blood directly from the warmest part of the shark. Before
collecting in the orbital sinus, this red muscle vein bathes the brain in warm
blood. These morphological data suggest the salmon shark has a significant
capacity to warm the brain and eyes.
INTRODUCTION
Temperature measurements have shown that the salmon shark (Lamna
ditropis, Hubbs and Follet, 1947) is one of very few fish and only seven sharks
able to elevate its body temperature above that of the surrounding water.
Differences of 8 to 11 °C between body core and ambient waters have been
measured (Smith, 1983), lending credence to the theory that the extent of the
red muscle, muscle retia, and supra-hepatic rete indicate degree of endothermy
(Carey et al., 1985). Recent research has confirmed these anatomical findings
and additionally identified an extensive kidney rete (Condie, unpublished).
These retia minimize the heat losses associated with breathing through
water. Since water has 1/40th the oxygen content and 30 times the heat
capacity of air, blood sent to the gills to be oxygenated also reaches thermal
equilibrium with the water (Carey, 1982). Retia function as virtual thermal
barriers, transferring heat away from unoxygenated blood before it is cooled in
the gills. Counter-current heat exchange maximizes efficiency to near 97% (for
the bluefin tuna, Carey, 1982), allowing fish like L. ditropis to maintain
observed temperature elevations. The accompanying advantages are potentially
great. The salmon shark almost certainly benefits from an expanded thermal
niche and increased aerobic capacity (Block & Finnerty, 1994).
No temperature measurements have been taken in the cranial region of
the salmon shark, but there is reason to believe fish would warm their brains
and eyes as well. Critical functions including nerve transmission, learned
behavior, and visual acuity are likely affected by temperature (Linthicum &
Carey, 1972). In addition, specific anatomical features for both preserving and
contributing heat have been associated with endothermy of the brain and eye.
It is hypothesized that determining the presence of these key morphological
features will afford an estimate of the salmon shark’s cranial endothermy.
MATERIALS AND METHODS
Three salmon sharks were obtained for study, hereafter referred to as
LDI, LD2, and LD4. LD1 and LD2 were captured in gill nets by the Monterey
Fishing Company, while LD4 was found beached in Monterey Bay. All had
localized handling wounds but were otherwise in excellent condition, showing
no damage from -20 °C storage. Morphometrics were taken and pertinent
information follows.
Weight
Sex
Fork Length
Specimen
92 cm
LDI
Female
34.9 kg
45.2
144
LD2
Male
129
LD4
Female
12.1
Latex Injections
Radio-opaque latex was used as an aid in dissection and visualization.
All injection was accomplished using Microfil Injection Compounds MV-120
(blue) and MV-130 (red). The suggested component ratios were modified
slightly to extend working time past ½ hour.
In order to inject the arterial system of the cranium, a window was
opened large enough to expose the pseudobranchial artery where it is closest to
the skin — at the point where it passes laterally over the hyomandibular
cartilage (Fig. 1). For LD1, 8 mL of latex were sufficient to inject the entire
cranium from one side. (Volumes exclude losses in the syringe and catheter,
but may include leakage.) LD2 and LD4 required injection from both sides,
requiring 8.5 and 3 mL for the former and 7.2 and 1.8 mL for the latter.
The venous system was injected from the origin of one of the two red
muscle veins in LD2 and LD4, using 38 and 8 mL of latex, respectively. The
second red muscle vein remained uninjected in LD2 and was back-injected
with 2 mL of latex at the point of its entry into the spinal cord.
Radiographs
The heads of LD2 and LD4 were x-rayed after decapitation. LD2 was x-
rayed again after removal of the jaws, and again after removal of excess tissue.
Vessel Anatomy/Diameters
Each head was dissected extensively to follow all vessels of interest. One
orbital rete was extracted from each specimen and sectioned. Vessel diameters
were measured both superficially and in cross section using an optical
micrometer.
Diameters of the red muscle vein were obtained from cross sections. To
estimate the relative amount of red muscle at a given body length, the red
muscle of successive steaks was isolated and weighed, adjusting for width of
the steak.
RESULTS

Cranial Anatomy
Cranial anatomy was found to be tailored to the presence of an orbital
rete on each side of the cranium. (Refer to Figure 2 for a diagrammatic
representation of the arterial system.) Both arterial contributors to the rete
arise from the first efferent branchial artery. At its origin, the pseudobranchial
artery is at least as large as this source. After passing laterally over the
hyomandibular cartilage and then between the cranial wall and the quadrate
process of the jaw, an average diameter above 2 mm was measured. (Vessel
diameters reported here are a consensus from the three specimens. See Figure 3
for individual results) No curvature is apparent until the vessel has reached the
spiracular pocket, where it successively branches into five to seven vessels that
curve dorsally, forming the pseudobranch plexus. These 1.5 mm vessels
quickly return ventrally, converge in an analogous fashion, and continue as one
vessel of diameter 2.5 mm. After approximately four large curves, this vessel
divides and begins the orbital rete, the cohesive meshwork of winding
arterioles lying within the orbital sinus (Fig. 4). Proceeding anteriorly, the
number of vessels approaches 100, and the diameters progressively decrease to
0.4 mm. Within a given cross section, vessels show little variation in size and
are evenly distributed, except at the beginning of the rete where diameters are
large.
The second tributary of the orbital rete begins as the efferent hyoidean,
arising from the efferent branchial artery below the pseudobranchial origin.
This artery travels across the basal plate in small curves, occasionally doubling
back on itself, with a relatively constant diameter near or below 1.5 mm. At the
orbital fenestra it divides. The internal carotid passes through the cranium into
the brain, while the stapedial branch passes into the fenestra and quickly
divides. The rete it forms lies between the pseudobranchial portion of the
orbital rete and the cranial wall. It is similar to, but smaller than its neighbor,
with vessel diameters decreasing to 0.35 mm at the anterior end.
Small arteries leave the rete for the eye, eye muscle, and mandible, but
the majority of blood (estimated by cross-sectional vessel area) leaves via a
gathering of arterioles at the far anterior end of the rete. These become the
cerebral arteries that, along with the internal carotid, supply the brain.
Ventro-dorsal and lateral x-rays of the brain precisely illustrate the extent
of the orbital rete, since the surrounding tissue remains undisturbed
(Figs. 5 & 6). The largest rete measures 9.5 cm from the beginning of the
pseudobranchial plexus to the anterior end. Figure 7 contains additional
measurements from the radiographs.
Vessels within the orbital rete are bathed in venous blood. That is,
arterial blood is only separated from venous blood by the media and adventitia
of the arterial walls, though they are relatively thick. A sparse web of tissue
connects arterial walls to one another or to the thin mesentery that
encapsulates the rete (Fig. 8).
The orbital rete lies completely within the orbital sinus (Block, personal
observation). This cavity is of greatest size near the eye and extends posteriorly
beside the rete. to eventually join the jugular vein (Burne, 1923). Venous
injection of the sinus via the myelonal vein was unsatisfactory.
Red Muscle Vein Anatomy
The red muscle vein was found to begin posteriorly only after the red
muscle had reached a significant size, and to enter the neural canal anteriorly
before the red muscle tapered away — at about the level of the last gill slit (Fig
9). Its diameter steadily increases to about 3.5 mm just before entering the
neural canal and joining with the myelonal vein. This vein follows the neural
canal into the cranium, at which point it joins the net of veins covering the
brain and finally drains into the orbital sinus.
DISCUSSION
Previous examination of 18 species of sharks has shown that warm
brains and eyes occur only in sharks that have an orbital rete (Block & Carey,
1985). The presence of a well developed orbital rete in Lamna ditropis therefore
strongly suggests that it is warming its brain and eyes. The rete is positioned
to intercept cool, oxygenated blood from the gills before it is distributed in
the cranium. In the rete, the high surface area afforded by the dividing and
coiling arterioles is well suited for counter-current heat exchange. Venous
blood flowing in the opposite direction is unencumbered by vessel walls,
bounded only by the mesentery encircling the rete. By having this stream of
warm venous blood pass directly over the arterioles while a separate stream fills
the remainder of the orbital sinus, a system of double insulation is
approximated. Isolation and efficiency of heat exchange may be increased by
this arrangement.
A previous study found lamnid sharks to have a range of vessel
diameters for the orbital rete from 600-800um to 150-300um, posterior to
anterior. The vessels of one 114 cm porbeagle shark (Lamna nasus) ranged
from 750 um to 300 um in the pseudobranchial plexus (Block & Carey, 1985).
Corresponding diameters for the three salmon shark specimens were generally
larger, at least at the posterior end, with LD1 diameters ranging from 1220 +
310um to 300 + 70um, LD2 from 2180 + 650um to 490 + 50um, and LD4
from 1220 + 260 to 300 + 120. Though LD1 and LD2 were, larger sharks than
the porbeagle (15 and 30 cm longer, respectively), LD4 was 22 cm shorter.
It is customary to link larger vessel size with diminished capacity for
heat exchange. However, other variables such as overall rete size or vessel wall
diameter were not taken into consideration. Additionally, there is a tradeoff
between heat exchange and unwanted oxygen exchange that depends on vessel
diameter. (Carey, 1982). Finding an optimum balance for heat exchange may
depend on unmeasured factors such as blood velocity or possible temperature
effects on hemoglobin (Carey, 1982).
The cranium can receive heat from a variety of locations. The red color
of the eye muscle has been noted by Wolf et al. as an indication of increased
oxidative fibers. These aerobic eye muscles probably contribute heat directly to
the orbital sinus in which they lie (Wolf et al., 1988). Nearby jaw muscle was
also observed to be red and of considerable size, perhaps helping warm the
general area. The brain itself is metabolically active, potentially heat-producing,
and drained by the cerebral veins which lead to the sinus. Finally, the red
muscle vein of L. ditropis may add a degree of endothermy to the cranium by
bringing heat directly from the warmest region of the body—the swimming
musculature (Wolf, et al. 1988). There is little reason, other than to provide
heat, for the venous net surrounding the brain to be of such significant size. It
was confirmed that the red muscle vein exists only within regions of
considerable red muscle, supporting its proposed purpose of transporting
thermal energy only from the warmest region of the body.
It will not be long before morphometric information on L. ditropis will
be supplemented by cranial temperature readings and information on the
extent of ambient thermal preferences. Three Pop-Up-Tags will be deployed
on salmon sharks in Alaska in June of 1997.
CONCLUSTONS
It is believed that the cranial, orbital rete, and red muscle vein anatomy
of the three salmon shark specimens, in combination with previous reports of
a warm body, are more than sufficient to indicate that Lamna ditropis has a
high degree of endothermy in the brain and eye.
LITERATURE CITED
Block, B.A., J.R. Finnerty. 1994. Endothermy in fishes: a phylogenetic analysis of
constraints, predispositions, and selection pressures. Experimental Biology of Fishes.
40: 283-302.
Block, B.A. & F.G. Carey. 1985. Warm brain and eye temperatures in sharks. J. Comp.
Physiol. B. 156: 229-236
Burne, R.H. 1923. Some peculiarities of the blood vascular system of the porbeagle shark
(Lamna cornubica). Philos. Trans. R. Soc. London. 212B: 209-257.
Carey F. G. 1982 Warm Fish. Pp 216-233 in A companion to animal physiology (C.R.
Taylor, K. Johansen, and L. Bolis, eds.), Cambridge Univ. Press. 365 pp.
Carey, F.G., J.G. Casey, H.L. Pratt, D. Urkuhart & J.E. McCosker. 1985. Temperature,
heat production, and heat exchange in lamnid sharks. Memoirs of the Southern
California Academy of Science 9: 92-108.
Condie, Jana Elizabeth. Anatomical adaptations for endothermy: retia mirabilia in the
salmon shark, Lamna ditropis. Harold Miller Library, Hopkins Marine Station,
Stanford University.
Linthicum, D.S. & F.G. Carey. 1972. Regulation of the brain and eye temperatures by the
bluefin tuna. Comp. Biochem. Physiol. 43A: 425-433.
Smith, R.L., D. Rhodes. 1983. Short note: body temperature of the salmon shark,
Lamna ditropis. J. mar. biol. Ass. U.K. 63: 243-244.
Wolf, N.G.. P.R. Swift & F.G. Carey. 1988. Swimming muscle helps warm the brain of
lamnid sharks. J. Comp. Physiol. B. 157: 709-715.
FIGURE LEGENDS
Figure 1: Top: LD2 before dissection. Bottom: Injection of the right
pseudobranchial artery with red latex.
Figure 2: Arterial system of the L. ditropis cranium: Diagrammatic outline of
the arterial system as it pertains to the orbital rete. Note that colors
designate temperature, not arteries and veins. Blood from the efferent
branchial arteries enters the orbital rete via the pseudobranchial artery and
the anterior carotid artery. There it is warmed by counter-current heat
exchange with venous blood (not shown).
Figure 3: Diameters of the stapedial and pseudobranchial plexus of the orbital
rete, posterior to anterior. Note that a decrease in vessel size is associated
with an increase in the number of arteries. Error bars denote standard
deviation of the sampled vessels. In areas with only a single cross-sectional
vessel, the S.D. reflects repeated measurements of the same vessel within one
cm.
Figure 4: Top: Dissection exposing right orbital rete of LD2. Bottom:
Doubly-injected left orbital rete extracted from LD2. PP pseudobranchial
plexus; SP stapedial plexus; PA pseudobranchial artery; OA optic artery; AE
arterial exit to anterior and posterior cerebral arteries; VE venous entry from
cerebral veins.
Figure 5: Radiograph of LD2 with jaws and excess tissue removed. Note
prominent left and right pseudobranchial plexus of the orbital retia.
10
Figure 6: Contact-printed radiograph of LD4. Jaws in place. PP
pseudobranchial plexus; SP stapedial plexus.
Figure 7: Pertinent measurements from LD2 and LD4 radiographs.
Figure 8: Top: Cross-sections from LD4 pseudobranchial plexus, posterior
(left) to anterior (right). Bottom: close-up of LD2 cross-section, showing
thick arterial walls connected by a sparse net of tissue. Venous blood (blue
latex) is unencumbered by vessel walls.
Figure 9: Diameters of the red muscle vein before and after entering the neural
canal of the spinal cord. S.D. bars reflect repeated measurements of the vein
within one cm. Amount of red muscle is an arbitrarily scaled measure of the
quantity of red muscle at the corresponding point where the vessel diameter
was taken.
11
ACKNOWLEDGMENTS
Numerous people have contributed immensely to this Hopkins research
project. I sincerely thank my advisor, Barbara Block, for her enthusiasm,
expertise, and confidence that I could fit 50 slides into a 15 minute talk. Beth
Condie was an incredible lab partner and I was lucky to have her help. I would
like to thank Dr. Tom Williams and his veterinary staff for allowing me to take
x-rays of an unlikely pet. I greatly appreciate the additional help of Henry
Mollet, Chuck Farwell, Carol Reeb, Ellen Freund, Heidi Dewar, Chris Patton,
and the Hopkins Marine Station staff.
12
Fig. 1
O

Fig. 2
Fig. 3
Orbital Rete Vessel Diameter
Stapedial Plexus
2-

+ 1.5.
0.


0.5.
O 3
0-

Stapedial
Plexus





7



Pseudobranchial
Plexus
Pseudobranchial Plexus
3.5.
3.5
X—— LD 1
1
3
LD2
2544
- 2.5
O LD 4
O.
8 9

8
5 158
1.5
O0.
8

.
8
4
+k
8 0 9 0 0 44
0.5.
4 0.5
0
Posterior
Anterior
Relative Position along Rete
- 1.5
0.5
0
Fig. 4
Fig. 5
Fig. 6
Orbital Rete Measurements
from LD2 and LD4 Radio
graphs
Left
Right
LD2: 1.9
LD4: 1.2
1D2.23
104.17
LD2:1.8
LD4: 0.8
Pseudobranchial
Plexus
Stapedial
Plexus
LD2: 9.5
LD4: 1.9
D4. 6.5
104.40
LD4:3.8
+
LD2:5.2
Fig. 7
LD2: 8.0
LD4:5.6

—
O
.
o
o





-

Arbitrary units
L


O
Q
koo
(uu) 2zi ss

5

8
Fig. 9