JOURNAL OF MORPHOLOGY 255:44 –57 (2003)
Development of the Supraorbital and Mandibular Lateral
Line Canals in the Cichlid, Archocentrus nigrofasciatus
Melissa L. Tarby and Jacqueline F. Webb*
Department of Biology, Villanova University, Villanova, Pennsylvania 19085
ABSTRACT The development of two of the cranial lat-
eral line canals is described in the cichlid, Archocentrus
nigrofasciatus. Four stages of canal morphogenesis are
defined based on histological analysis of the supraorbital
and mandibular canals. “Canal enclosure” and “canal os-
sification” are defined as two discrete stages in lateral line
canal development, which differ in duration, an observa-
tion that has interesting implications for the ontogeny of
lateral line function. Canal diameter in the vicinity of
individual neuromasts begins to increase before ossifica-
tion of the canal roof in each canal segment; this increase
in canal diameter is accompanied by an increase in canal
neuromast size. The mandibular canal generally develops
later than the supraorbital canal in this species, but in
both of these canals development of the different canal
segments contained within a single dermal bone is asyn-
chronous. These observations suggest that a dynamic pro-
cess requiring integration and interaction among different
tissues, in both space and time, underlies the development
of the cranial lateral line canal system. The supraorbital
and mandibular canals appear to demonstrate a “one-
component” pattern of development in Archocentrus nigro-
fasciatus, where the walls of each canal segment grow up
from the underlying dermal bone and then fuse to form
the bony canal roof. This is contrary to numerous pub-
lished reports that describe a “two-component” pattern of
development in teleosts where the bony canal ossifies sep-
arately and then fuses with an underlying dermal bone. A
survey of the literature in which lateral line canal devel-
opment is described using histological analysis suggests
that the occurrence of two different patterns of canal mor-
phogenesis (“one-component” and “two-component”) may
be due to phylogenetic variation in the pattern of the
development of the lateral line canals. J. Morphol. 255:
44 –57, 2003. © 2002 Wiley-Liss, Inc.
KEY WORDS: osteology; lateral line; neuromasts; Cichli-
dae; Archocentrus; dermal bone
The mechanosensory lateral line system is a prim-
itive vertebrate sensory system composed of a series
of hair cell-based neuromast receptor organs located
on the skin and in ossified lateral line canals. These
canals are prominent features of the dermatocra-
nium of bony fishes; they are found in a well-defined
subset of dermal cranial bones in all major osteich-
thyan lineages (Webb, 1989b). It has been suggested
that the canal-bearing bones of the skull are com-
posite structures, which are the result of the fusion
of two components: an ossified, tubular lateral line
© 2002 WILEY-LISS, INC.
canal and an underlying dermatocranial bone. This
idea was embraced by most late 19th and early 20th
century investigators who used histological analysis
to study the morphology and development of the
lateral line canal system and the dermal cranial
bones in a variety of taxa (see Pehrson, 1922; Dev-
illers, 1947; Stensio, 1947; Kapoor, 1970; Jollie,
1984a,b). Despite the popularity of this concept, the
terminology used to define the two components var-
ied widely and has resulted in some confusion. The
tubular, lateral line component has been called the
“lateral ossification” (Swertzoff, 1926 in Kapoor,
1970), “sensory-line component” (Stensio, 1947),
“latero-sensory component” (Lekander, 1949), and
the “neurodermal component” (Sire and Huysseune,
1993). Similarly, the underlying dermal bone has
been described as the “medial ossification” (Swert-
zoff, 1926, in Kapoor, 1970), “anamestic component”
(Stensio, 1947), “membranous component” (Lek-
ander, 1949), “lamellar component” (Lekander,
1949), and the “membranodermal component” (Sire
and Huysseune, 1993).
The development of lateral line bones from two
bony components has been documented in Amia (Al-
lis, 1889; Pehrson, 1922, 1940) and has been sug-
gested by features of the adult osteology of Lepisos-
teus (Aumonier, 1941; deBeer, 1985), elopomorphs
(Muraenidae, Allis, 1903; Anguilla, de Beer, 1985),
esocids (Esox, Pehrson, 1944), several ostariophys-
ans (Kindred, 1919; Lekander, 1949; Kapoor, 1961;
Reno, 1966) and salmonids (Jollie, 1984b; deBeer,
1985, reviewed in Pehrson, 1922). In ostariophysan
fishes, for example, the laterosensory component
and the underlying lamellar component develop
from two separate centers of intramembranous os-
sification, which subsequently fuse (e.g., the cyp-
rinid, Phoxinus phoxinus, Lekander, 1949). In other
ostariophysans, the two components appear to be in
such close proximity that they appear to develop
*Correspondence to: Jacqueline F. Webb, Department of Biology,
Villanova University, Villanova, PA 19085.
E-mail: jacqueline.webb@villanova.edu
Published online 00 Month 2002 in
Wiley InterScience (www.interscience.wiley.com)
DOI: 10.1002/jmor.10045
 CRANIAL LATERAL LINE CANAL DEVELOPMENT
from one center of ossification (e.g., Leuciscus ruti-
lus, Lekander, 1949). Lekander (1949) showed that
in some portions of the canal system in the ostari-
ophysan Nemachilus barbatula both components
are present but that the laterosensory component
fails to fuse with the lamellar component, further
supporting the hypothesis that the canal and under-
lying bone develop independently. This sort of mor-
phology has also been documented in fossil acti-
nopterygians (Pehrson, 1922). Furthermore, the
occasional absence of one of the two components
further supports the idea that the two components of
lateral line bones develop separately. For instance,
in Cobitis taenia, an ostariophysan with a reduced
canal pattern (where canals containing neuromasts
are replaced by superficial neuromasts in the epithe-
lium, Webb, 1989b), the canal component is absent,
but the underlying lamellar bone is present in some
canal segments (Lekander, 1949; see discussion in
Kapoor, 1970).
The fact that lateral line canals become integrated
in dermal bones during ontogeny requires that there
be precise interactions in both space and time be-
tween presumptive canal neuromasts (superficial
neuromasts that will later become enclosed in a
canal) or the bony lateral line canal and the under-
lying dermal bone. But despite a growing interest in
the development of the vertebrate skull (see Hall
and Hanken, 1985; Hanken and Thorogood, 1993),
the development of the lateral line canal system in
fishes has not received much attention (but see
Kapoor, 1970, Jollie, 1984a,b).
Recent studies of the functional development of
fish sensory systems document the size and distri-
bution of superficial neuromasts and the general
timing of canal appearance (e.g., Neave, 1986; Pol-
ing and Fuiman, 1997), but do not provide detailed
descriptions of the pattern of lateral line canal mor-
phogenesis. Recent comprehensive studies of the
postembryonic development of the skull of various
teleosts have used enzymatically cleared and
stained osteological preparations and have docu-
mented the appearance of ossified canal walls and
canal roofs and the number and position of canal
pores (Amphistichus, Morris and Gaudin, 1982; cen-
trarchids, Mabee, 1993; Danio, Cubbage and Mabee,
1996; Betta, Mabee and Trendler, 1996; Clarias, Ad-
riaens et al., 1997; Scophthalmus, Wagemans et al.,
1998), but do not address the issue of how lateral
line canals become integrated into dermal cranial
bones. Mabee (1993) however, did note that “perpen-
dicular ridges” (⫽ canal walls) appear to be “contin-
uous membranous extensions” of dermal bones in
cleared and stained centrarchid fishes, suggesting a
one-component pattern of lateral line bone develop-
ment in these perciform fishes.
The purpose of this study was to describe the
development of the lateral line canals in the cichlid
Archocentrus nigrofasciatus (⫽ Cichlasoma nigro-
fasciatum, e.g., Webb, 1989c) in order to test the
45
hypothesis that the integration of the lateral line
canals into dermal cranial bones is the result of the
fusion of two components, a lateral line canal and an
underlying dermal cranial bone. Histological mate-
rial, cleared and stained specimens, and material
prepared for scanning electron microscopy were
used to describe the pattern of postembryonic devel-
opment of the supraorbital and mandibular lateral
line canals. These canals were chosen for detailed
analysis because they are located in a small number
of well-ossified and easily identifiable skeletal ele-
ments (the nasal, frontal, dentary, and anguloar-
ticular) that ossify relatively early. Because they are
rostro– caudal in orientation, they are conducive to
histological analysis of transverse series. This study
will serve as a baseline for further examination of
the patterns of cranial lateral line canal develop-
ment in teleosts and the developmental mechanisms
that underlie them.
MATERIALS AND METHODS
Archocentrus nigrofasciatus, the convict cichlid, is a Central
American cichlid (Axelrod and Vorderwinkler, 1980) which is
easily obtained through commercial suppliers and is easy to breed
in the lab. The pink morph has little or no dark pigment, which
could otherwise interfere with the interpretation of histological
data. In addition, this species has a relatively unspecialized,
narrow lateral line canal system (Webb, 1989b) and dermal cra-
nial skeleton and can serve as a model for the study of patterns
and mechanisms of lateral line canal development.
Fish were purchased from a local pet supplier and were main-
tained in 20-gallon aquaria equipped with undergravel filters and
heaters at a mean daily temperature of 23.3–24.7°C and were
provided with a 12-h light/dark cycle (Tarby, 1998). Adults were
fed daily with commercial flake food. One male and several fe-
males were placed in a tank and allowed to form pair bonds.
Several days to several weeks later, when a female laid eggs in a
small clay pot provided for the purpose, the other females were
removed. The parents were allowed to protect the eggs for 2 days
and were then separated from them. The clay pot was aerated
with an air stone (as a substitute for the ventilation normally
provided by parents) for at least 4 days, or until the fry hatched
and began to swim freely. Hatching occurred 4 days post–
fertilization and fry swam freely 4 days later. Fry were fed com-
mercial liquid fry food 2– 4 times a day until approximately 10
days post– hatch, after which they were fed flake food daily.
One male and two females served as parents of all individuals
in the three broods used in this study. Individuals from Broods A
and C were full siblings, while individuals from Brood B were
half-siblings of the individuals in Broods A and C (sharing the
same father). Starting from the day of hatch, three individuals
from Broods A and B were sampled every 3 days and four indi-
viduals from Brood C were sampled daily. All fishes were anes-
thetized in MS 222 in tank water until unresponsive. Fishes from
Broods A and B were fixed and stored in chilled (4°C) 10%
phosphate-buffered formalin (pH 7.4). Fishes from Brood C were
fixed and stored in chilled (4°C) 10% formalin in PBS (0.1M
phosphate buffer, 0.9% NaCl, pH 7.4). Munz (1986) demonstrated
that the development of the lateral line system is correlated with
size, not age. Thus, the notochordal length or standard length
(SL) of all fixed individuals was measured to the nearest 0.5 mm
within 1 week after fixation. Flexion is complete and the hypural
plate is evident at 7 days post– hatch (5 mm SL). Fishes increased
in size at an overall rate of between 0.12– 0.16 mm/day (Fig. 1).
46 M.L. TARBY AND J.F. WEBB
Fig. 1. Size of Archocentrus nigrofasciatus vs. age (days post-
hatch). Brood A was used for cleared and stained material (y ⫽
4.409 ⫹ 0.161x; P ⬍ 0.0001; r2 ⫽ 0.93). Brood B was used for
cleared and stained material (y ⫽ 4.644 ⫹ 0.154x; P ⬍ 0.0001;
r2 ⫽ 0.90). Brood C was used for paraffin and plastic resin his-
tology, and for SEM (y ⫽ 3.839 ⫹ 0.124x; P ⬍ 0.0001; r2 ⫽ 0.82).
Clearing and Staining
Ontogenetic series from Broods A (n ⫽ 24, 4.5–15.5 mm SL, 0.5
mm intervals, one individual/size class) and Brood B (n ⫽ 15,
4.0 –12.5 mm SL, 0.5 mm intervals, one individual/size class)
were cleared and stained for bone (alizarin red) and cartilage
(Alcian blue) following Pothoff (1983). This material was used to
determine the relative timing of the ossification of the cranial
bones and the size range needed for a subsequent histological
analysis of canal morphogenesis. Two older, larger specimens
from Brood B (33 and 41 mm SL) were cleared and stained and
used to illustrate adult morphology. A complete developmental
osteology was not attempted.
Scanning Electron Microscopy
Six individuals from Brood C (4.5– 8.0 mm SL) were prepared
for SEM in order to identify and distinguish between presumptive
canal neuromasts and other superficial neuromasts (those that
remain superficial in adults) and to confirm the order and timing
of canal enclosure observed in histological material. Fishes were
dehydrated in an ascending ethanol series, critical point-dried in
CO2, mounted on stubs with silver paint, sputter-coated with
gold-palladium alloy (following Webb, 1989c), and viewed using a
Hitachi model S5-7 scanning electron microscope.
Histology
Fifteen individuals from Brood C (4.5–11.5 mm SL, 0.5 mm
intervals, one individual/size class) were used to analyze the
pattern and relative timing of the ossification of each canal seg-
ment in the supraorbital and mandibular canals. One to 14 days
after fixation, these fishes were decalcified for 4 h (Cal-Ex, Fisher,
Orangeburg, NY), rinsed in PBS, transferred through a graded
sucrose series, dehydrated in a graded ethanol/t-butanol series,
and embedded in Paraplast Plus (Fisher). Blocks were sectioned
transversely at 10 ␮m and stained with the HBQ stain for bone,
cartilage, and connective tissue (Hall, 1986), which was modified
to include an additional 2-min acid alcohol rinse following May-
er’s hematoxylin in order to enhance nuclear staining. All neuro-
masts in the supraorbital and mandibular canals were identified
and four developmental stages that describe the pattern of devel-
opment of the lateral line canal in the vicinity of individual
neuromasts were defined.
Seven additional individuals from Brood C (4.5– 8.0 mm SL, at
0.5-mm intervals) were prepared for plastic resin histology to
provide increased resolution for the identification of neuromasts
and sites of intramembranous ossification. These fishes were
decalcified for 3 h (Cal-Ex, Fisher), rinsed in PBS, dehydrated in
a graded ethanol series, infiltrated overnight in Historesin (Leica,
Deerfield, IL), and embedded. Tissue was sectioned at 5 ␮m,
using a tungsten carbide knife on a Leica motorized microtome.
Alternate sections were mounted on clean glass slides, yielding a
10 ␮m inter–section interval. Slides were dried overnight at 60°C,
stained with 0.5% cresyl violet, differentiated in running tap
water, air-dried overnight, and coverslipped.
Paraffin-embedded histological series were also used to evalu-
ate ontogenetic trends in neuromast size and in canal diameter in
both the supraorbital and mandibular canals. In fishes with both
narrow and wide canal systems (defined by Webb, 1989b), canal
neuromasts span the width of the canals (Coombs et al., 1988;
Webb, 1989a). Since neuromast size is correlated with fish size
among teleosts, and neuromast size is correlated with the diam-
eter of the canal in which they are located in adult fishes (Münz,
1989), it follows that there should be a positive correlation be-
tween canal diameter and neuromast size during ontogeny. Neu-
romast width was measured at the rostrocaudal midpoint of each
neuromast. Canal diameter was measured parallel to the surface
of each neuromast in the same sections in which neuromast width
was measured. All measurements were made using an ocular
micrometer. Least-squares analysis (MS Excel 98 and JMP 3.1.5)
and geometric mean functional regression analysis (Ricker, 1973)
were used to describe the relationship between fish size and canal
diameter and between canal diameter and neuromast width
(where there is variability on both axes) in the vicinity of each of
the supraorbital and mandibular canal neuromasts in histologi-
cal series. Since Archocentrus nigrofasciatus has a narrow canal
system, which is defined as having a relatively uniform canal
diameter (Webb, 1989b), canal diameter in the vicinity of canal
neuromasts was considered representative of the canal diameter
throughout the cranial lateral line canal system.
RESULTS
The supraorbital (SO) canal is contained in the
nasal bone and in the frontal bone, which forms the
dorsal roof of the neurocranium medial to the eyes.
The SO canal contains five neuromasts identified as
SO1–5 in rostral to caudal sequence; one neuromast
is located between adjacent canal pores. Neuromast
SO1 is in the portion of the SO canal in the nasal
bone, while neuromasts SO2–5 are in the portion of
the SO canal in the frontal bone. The lumen of the
SO canal is contiguous with the canal in the pterotic
(Fig. 2A).
Neuromast SO1 is found inside the tubular na-
sal bone (Fig. 2A). The pore rostral to this neuro-
mast is defined as the anterior terminal pore of
this canal segment. The pore caudal to this neu-
romast is adjacent to the terminal pore of the
rostral–most segment of the SO canal, which is
contained in the frontal bone. A common epithelial
pore connecting the canal lumen to the external
environment is present at this location, but the
two bony canal segments (which are in adjacent
 CRANIAL LATERAL LINE CANAL DEVELOPMENT
Fig. 2. Camera lucida drawing of cleared and stained Ar-
chocentrus nigrofasciatus (33 mm SL) showing lateral line canals
and location of canal neuromasts (numbers) in the neurocranium
(supraorbital canal, SO) and the mandible (mandibular canal,
MD). A: Dorsal view indicating the location of the supraorbital
canals (SO) and canal neuromasts contained in the nasal and
frontal bones (numbers ⫽ neuromasts SO1–5). B: Ventrolateral
view of the mandible indicating the location of the mandibular
canal and canal neuromasts (numbers ⫽ neuromasts MD1– 4) in
the canal portions in the dentary and anguloarticular bones. aa,
anguloarticular; de, dentary; fr, frontal; na, nasal.
dermal bones) are not fused. Neuromast SO2 is
located in the canal in the frontal bone just caudal
to this terminal pore. Neuromasts SO2–SO5 are
found between sequential pores along the canal.
The pore between SO3 and SO4 is unusual in that
it forms a common medial pore with its homolog on
the other side of the head, linking the right and
left SO canals (Fig. 2A). The posterior terminal
pore of the SO canal is located caudal to SO5, at
the articulation of the frontal bone and the
pterotic just caudal to it.
The mandibular canal (MD) is located within
the dentary and anguloarticular bones and con-
tains four neuromasts, MD1– 4 in rostral-caudal
sequence (Fig. 2B); one neuromast is located be-
tween adjacent canal pores. Neuromasts MD1–3
are located in the portion of the canal in the den-
tary bone, while neuromast MD4 is located in the
portion of the canal in the anguloarticular bone.
The anterior terminal pore of the MD canal is
located at the anterior end of the mandible. The
posterior terminal pore of the canal portion in the
dentary is at the posterior end of the dentary bone
where the canal continues into the anguloarticu-
lar bone. These two bones are not fused; a common
epithelial pore joins the posterior terminal pore of
47
the portion of the canal in the dentary and the
anterior terminal pore of the portion of the canal
in the anguloarticular. The posterior terminal
pore of the canal portion in the anguloarticular
defines the posterior end of the mandibular canal,
which is contiguous with the preopercular canal.
Brief Overview of Cranial Development in
Archocentrus nigrofasciatus
In the smallest cleared and stained individuals
examined (4.5 mm SL), cranial ossification is limited
to the mandible, maxilla, and premaxilla. In a
5.5-mm individual, the majority of the mandible,
maxilla, and premaxilla are ossified. A 6.0-mm indi-
vidual shows partial ossification of the frontal bone,
but the canals in the frontal bone are not ossified
and are therefore not visible. Short ossified canal
walls of the portion of the supraorbital canal in the
frontal bone (SO2–5) are apparent in a 6.5-mm in-
dividual. In a 7.5-mm SL individual, most of the
neurocranium appears to be ossified and the walls of
the SO canal are visible. Canal pores are present in
an 8.5-mm individual, indicating that the canal roof
of the SO canal has ossified. In all individuals larger
than 8.5 mm, the SO canal pores are clearly visible,
indicating that at least some canal roof ossification
is complete. Mandibular ossification is apparent in
individuals ⬎4.5 mm SL. Mandibular canal walls
are ossified in a 7.5-mm individual and canal walls
begin arching to form the canal roof in 8.5-mm indi-
viduals. The ossified canal roof and some canal pores
of the mandibular canal are visible in individuals
⬎10.5 mm.
Fig. 3. Schematic representation of the four stages of canal
morphogenesis characteristic of the mandibular and supraorbital
canals in Archocentrus nigrofasciatus. A: Stage I: Superficial
neuromast (represented by hair cells with a kinocilium and
shorter stereocilia) located in epidermis (above the basement
membrane ⫽ black line) over the underlying dermal bone (stip-
pled). B: Stage II: Neuromast appears to be sunken into a groove
and ossified canal walls (stippled) extend upwards from the un-
derlying lamellar bone on either side of a neuromast. C: Stage III:
The epithelium has fused over the neuromast forming an epithe-
lial canal roof. D: Stage IV: The bony canal walls fuse over the
neuromast forming an ossified canal roof.
48 M.L. TARBY AND J.F. WEBB
Stages in the Development of the
Supraorbital and Mandibular Canals
The pattern of canal morphogenesis in the SO and
MD canals is defined as a series of four stages (Fig.
3). At Stage I (Fig. 3A), a flat bony plate is present in
the dermis beneath a presumptive canal neuromast,
which sits in the epithelium. At Stage II (Fig. 3B),
the neuromast appears to sit in a groove with two
parallel bony walls extending upward from the bony
plate. At Stage III (Fig. 3C), an epithelial canal
segment is present as the result of the fusion of the
edges of the epithelial groove covering the neuro-
mast; the canal walls continue to ossify intramem-
branously and begin to arch as they extend towards
the midline over the neuromast. At Stage IV (Fig.
3D), the ossified canal walls have fused over the
neuromast forming the bony canal roof. At this
point, morphogenesis of a canal segment around a
neuromast is considered complete. However, mor-
phogenesis of the canal (e.g., ossification of all of the
segments of the canal) continues as adjacent canal
segments fuse, leaving a common pore between
them.
Development of the Supraorbital Canal
In a 5.0-mm SL individual, only one presumptive
SO canal neuromast (neuromast SO5) is visible in
the epithelium. In histological section, the neuro-
masts in the vicinity of the future SO canal appear
as small, darkly staining clusters of cells. SEM con-
firmed that each cluster of cells is a small superficial
neuromast (Fig. 4A,B). Two superficial neuromasts
are present in a transverse row (first row) located
between the nares. A second transverse row of four
superficial neuromasts (second row) is present cau-
dal to the first row, and just medial and caudal to the
nares (Fig. 4B). In a 6.5-mm SL individual, both
transverse rows consist of four neuromasts (Fig.
4C,D). The lateral–most neuromasts in the second
row are presumptive canal neuromasts that sit in
grooves and represent the portion of the SO canal in
the nasal bones (neuromast SO1 of the right and left
sides); they appear elongated in the axis of the
groove in which they sit. This is typical of canal
neuromasts in narrow canals; superficial neuro-
masts remain circular in profile. After enclosure of
the SO canal, the two medial neuromasts in the first
and second rows remain superficial (8.0 mm SL, Fig.
4E,F). The five presumptive canal neuromasts of the
SO canal are illustrated in Figures 4A and 5A.
All five canal segments of the SO canal demon-
strate the same developmental pattern (Fig. 3),
but they develop asynchronously, with an overall
caudal to rostral sequence. Ossified bone repre-
senting the portion of the SO canal in the nasal
bone (SO1) is not apparent in individuals smaller
than 6.5 mm SL. In 6.5-mm (Figs. 4C,D, 5C) and
7.5-mm individuals (Fig. 6A,B), the canal walls on
either side of neuromast SO1 form a groove (Stage
II). In 8.0-mm (Figs. 4E, 5D) and 8.5-mm (Fig.
6C,D) individuals, the portion of the SO canal in
the nasal bone is enclosed by epithelium (Stage
III). In individuals ⬎9 mm SL, the canal segment
in the nasal bone is completely ossified (Stage IV,
Fig. 6E,F).
The segments of the SO canal in the vicinity of
neuromasts SO2–5, are all at Stage I in 5.0-mm and
5.5-mm individuals (Table 1). In a 6.0-mm individ-
ual, the canal segment in the vicinity of neuromast
SO2 is at Stage I, but those associated with neuro-
masts SO3–5 have advanced to Stage II (Fig. 5B).
Canal segments in the vicinity of neuromasts SO4
and SO5 are the first to enclose (Stage III) in a
6.5-mm individual (Table I, Figs. 5C, 7A), but the
canal segments in the vicinity of neuromasts SO2
and SO3 remain at Stage II. In a 7.5-mm individual,
all segments of the SO canal associated with the
frontal bone (SO2–5) are enclosed (Stage III). In an
8.0-mm individual, the canal roof over neuromast
SO5 is ossified (Stage IV) and in an 8.5-mm individ-
ual, the canal roof over neuromasts SO3 and SO4
are ossified. Finally, all SO neuromasts in the fron-
tal bone are surrounded by ossified canal segments
in individuals ⬎9.5 mm SL (Stage IV, Fig. 7B,C). In
summary, enclosure and ossification occur first in
the regions of neuromasts SO4 and SO5, then more
anteriorly in the frontal bone, in the vicinity of neu-
romasts SO2 and SO3. Once the portion of the SO
canal in the frontal bone has ossified, the canal
Fig. 4. Scanning electron micrographs of the dorsal aspect of
the anterior portion of the head of Archocentrus nigrofasciatus
illustrating development of the supraorbital canal (SO). Photos in
right column are enlargements of photos in left column. A: Pre-
sumptive canal neuromasts of the SO canal (numbers ⫽ SO1, 2,
3, 4, 5), (5.5 mm SL). B: Two superficial neuromasts (arrows) sit
rostral to the nasal epithelium (ol). Two transverse rows of su-
perficial neuromasts (1st and 2nd). The most lateral neuromasts
in the 2nd transverse row are presumptive canal neuromasts of
the supraorbital canal (SO 1, 2, and 3), (5.5 mm SL). C: The
lateralmost superficial neuromasts in the 2nd transverse row
(SO1) are in a groove, which represents the portion of the SO
canal in the nasal bone. The first neuromast in the portion of the
SO canal contained in the frontal bone (neuromast SO2) is also in
a groove. The portion of the SO canal containing neuromast SO3
is already enclosed (not illustrated). The pores caudal to neuro-
mast SO3 in the two bilateral SO canals are directed medially
(mp) and will eventually fuse to form a single pore in the dorsal
midline. The olfactory epithelia (ol) have sunk beneath the epi-
thelial surface, leaving a single naris (6.5 mm SL individual). D:
Closeup of image in C; noting neuromasts SO1 and SO2 in
grooves. E: Neuromast SO1 is enclosed within the canal segment
in the nasal bone. The two closely apposed pores between neuro-
masts SO1 and SO2 will fuse into a single pore later during
ontogeny (⬃8.0 mm SL). F: Closeup of E; noting the location of
canal segments containing SO1 and SO2. 1st, first transverse row
of superficial neuromasts; 2nd, second transverse row of superfi-
cial neuromasts; 1, 2, 3, SO canal neuromasts SO1, SO2, and SO3
(see Fig. 2); mp, pores that will fuse to form a common median
pore. ol, olfactory organ (or naris in D). For A,C,E, scale bar ⫽ 200
␮m; for B,D,F, scale bar ⫽ 100 ␮m.
CRANIAL LATERAL LINE CANAL DEVELOPMENT 49

Figure 4
50 M.L. TARBY AND J.F. WEBB

Fig. 5. Scanning electron

micrographs of a lateral view
of the head showing supraor-
bital canal (SO) development
in Archocentrus nigrofasciatus.
Numbers indicate position of
each of the five SO canal neu-
romasts (1–5 ⫽ SO1–5). A: 5.5
mm SL. Neuromasts SO3,
SO4, and SO5 sit in grooves. B:
6.0 mm SL: Canal enclosure
around SO4 has occurred and
is nearly complete around
SO5. Although not visible, SO3
remains in a deep groove,
while SO1 and SO2 are in shal-
low grooves. C: 6.5 mm SL:
Neuromasts SO3, SO4, and
SO5 are enclosed but canal
roofs are not ossified. Neuro-
masts SO1 and SO2 lie in
grooves. D: 8.0 mm SL: All SO
canal neuromasts are enclosed
and canal pores are clearly vis-
ible. Scale bars: A,B ⫽ 200␮m,
C,D ⫽ 300␮m.

segment in the nasal bone ossifies around neuro-
mast SO1.
Development of the Mandibular Canal
The mandibular canal (MD) is integrated into the
two dermal bones: the dentary and the anguloarticu-
lar. Like the segments of the SO canal, the MD canal
segments develop asynchronously. At 6.0 mm SL,
neuromasts MD1– 4 are superficial (Stage I; Table 1,
Fig. 8A) and sit in the epithelium overlying the
dentary bone. The canal segment around neuromast
MD3 is the first to develop and is at Stage II in a
6.5-mm individual (Fig. 8B). It is enclosed in a
7.5-mm individual (Stage III; Fig. 8C) and is ossified
in a 9.0-mm individual (Stage IV; Fig. 8D). The
canal segments around neuromasts MD2 and MD4
reach Stage II later, in a 9.5-mm individual, and the
segment around neuromast MD1 does not reach
Stage II until 10.5 mm. Neuromast MD2 is enclosed
at 10.5 mm and neuromasts MD1 and MD4 are
enclosed at 11 mm. The canal segment around neu-
romast MD2 is ossified at 11 mm and the segments
around neuromasts MD1 and MD4 are ossified in
individuals ⬎11.5 mm SL. Larger individuals in
which all MD canal segments were fully ossified
(Stage IV) were not available for study. To summa-
rize, the canal segment around neuromast MD3 de-
velops first, followed by the more rostral canal seg-
ment in the dentary (MD2). Portions of the canal in
the anteriormost end of the dentary (MD1) and in
the anguloarticular (MD4), at the caudal end of the
MD canal, ossify last, when fishes are ⬎11.5 mm SL.
Enclosure, Ossification, and Growth of the
Canals and Canal Neuromasts
In the SO canal, development of canal walls
(Stage II) begins at 6.0 mm SL; enclosure of canal
segments (Stage III) begins at 6.5 mm. Ossification
of canal segments (Stage IV) starts at 8.0 mm and
all canal segments are ossified at 9.0 mm (Table 1).
The mean growth interval (in mm SL) during which
individual SO canal segments 1–5 are at Stage II
(groove) is 0.8 mm (n ⫽ 5, range ⫽ 0.5–1.5 mm) and
the mean growth interval during which SO canal
segments are at Stage III (enclosed, with unossified
roof) is twice that, or 1.6 mm (n ⫽ 5, range ⫽ 1.0 –2.0
mm) (see Table 1).
MD canal development occurs later than SO
canal development (Table 1). Development of ca-
nal walls (Stage II) begins at 6.5 mm SL; enclosure
 CRANIAL LATERAL LINE CANAL DEVELOPMENT 51
TABLE 1. Order and timing of the development of the supraorbital and mandibular canals

SL SO1 SO2 SO3 SO4 SO5 MD1 MD2 MD3 MD4

5.0 I I I I I — — — —
5.5 I I I I I — — — —
6.0 I I II II II I I I I
6.5 II II II III III I I II I
7.0 II III III III III I I II I
7.5 II III III III III I I III I
8.0 III III III III IV I I III I
8.5 III III IV IV IV I I III I
9.0 IV IV IV IV IV I I IV I
9.5 IV IV IV IV IV I II IV II
10.0 IV IV IV IV IV I II IV II
10.5 IV IV IV IV IV II III IV II
11.0 IV IV IV IV IV III IV IV III
11.5 IV IV IV IV IV III IV IV III

Stages I–IV in the vicinity of supraorbital (SO1–5) and mandibular canal neuromasts (MD1– 4) based on histological material (n ⫽ 14,
5.0 –11.5 mm SL, one individual/size class). Bold indicates transition to next stage; —indicates that the developmental stage could not
be discerned.

TABLE 2. Canal diameter at enclosure (Stage III) and ossification (Stage IV) in the vicinity of individual canal neuromasts
in the supraorbital (SO1–5) and mandibular (MD1– 4) canals in Archocentrus nigrofasciatus

Enclosure (stage III) Ossification (stage IV)

Size (SL) Canal diam. (␮m) Size (SL) Canal diam. (␮m)

SO1 8.0 49 9.0 51

SO2 7.0 74 9.0 94
SO3 7.0 86 8.5 107
SO4 6.5 57 8.5 118
SO5 6.5 58 8.0 92
MD1 11.0 61 ⬎11.5 —
MD2 10.5 65 11.0 79
MD3 7.5 49 9.0 61
MD4 11.0 60 ⬎11.5 —

Based on Staging data in Table 1.

of neuromasts (Stage III) begins at 7.5 mm. Ossi-
fication of these two canal segments are complete
at 9.0 and 11 mm, respectively, but other seg-
ments do not ossify until fishes are ⬎11.5 mm. For
neuromast MD3, the growth interval (in mm SL)
during which the canal segment is at Stage II
(groove) is 1.0 mm and the growth interval during
which the canal segment is at stage III (enclosed,
with unossified roof) is 1.5 mm.
Canal diameter at the level of each neuromast in
the SO canal appears to increase linearly with fish
size subsequent to enclosure (Stage III) (Fig. 9A).
The segment of the SO canal in the nasal bone
encloses and ossifies at a smaller diameter (49 ␮m)
than the segments of the SO canal in the frontal
bone (57– 86 ␮m; Table 2). Ossification occurs later,
at a canal diameter of 51 ␮m in the nasal bone and
92–118 ␮m in the frontal bone. Thus, there may be
no change in the diameter of individual canal seg-
ments between initial enclosure and ossification of
the canal roof (SO1), or canal diameter may increase
two–fold with just a 2 mm increase in fish size (for
SO2–5). Neuromast width increases linearly with
canal diameter in the vicinity of each of the canal
neuromasts (Fig. 10A). The overall relationship be-
tween canal diameter and neuromast width for
SO1–5 shows a tight correlation (r2 ⫽ 0.83) with a
slope of 1.038, thus indicating that canal diameter
and neuromast width increase in concert as a fish
grows.
The portion of the MD canal around MD3, which
encloses and ossifies well before the other MD canal
segments, shows a linear increase in diameter sub-
sequent to enclosure and ossification (Fig. 9B). En-
closure in the MD canal occurs at a diameter of
49 – 65 ␮m; ossification occurs when canal diameter
is 61–79 ␮m (Table 2). The width of neuromast MD3,
which is enclosed early relative to the other MD
neuromasts, increases linearly as canal diameter
increases subsequent to enclosure (Stage III; Fig.
10B). The canal segments around MD1, 2, and 4 do
not show a linear increase in diameter subsequent to
enclosure and ossification, which occurs when fishes
are ⬎10.5 mm SL. As in the SO canal, there is a high
overall correlation between canal diameter and neu-
romast width (r2 ⫽ 0.74), with a slope of 1.355. This
slope, which appears to be higher than that for the
SO canal, indicates that the MD canal tends to be a
bit wider relative to neuromast width compared to
the SO canal in the same individuals.
52 M.L. TARBY AND J.F. WEBB

Fig. 6. Morphogenesis of the

portion of the supraorbital (SO)
canal in the nasal bone of Ar-
chocentrus nigrofasciatus, which
contains neuromast SO1. A: The
nasal bone is in the form of a
groove (Stage II, 7.5 mm SL). B:
Closeup of A; arrowheads denote
the distal ends of the canal walls
(stained pink). C: Canal is en-
closed around neuromast SO1
(Stage III) but the roof is not yet
ossified (8.5 mm SL). D: Closeup
of C; neuromast SO1 is visible,
arrowheads denote the distal
ends of the nasal bone ossifica-
tion, which extend towards the
epithelial canal roof. A superfi-
cial neuromast (sn) is in the ep-
ithelium, just medial to the ca-
nal. E: SO canal segment is
enclosed (11.5 mm SL, Stage IV).
F: Closeup of E showing ossifica-
tion of the canal roof (arrow-
heads, pink). Neuromast SO1 is
not visible in this section. A,C,E,
scale bars ⫽ 200 ␮m; B,D,F,
scale bars ⫽ 50 ␮m. oe, olfactory
epithelium; ol, olfactory organ;
nm, neuromast S1; sn, superfi-
cial neuromast; so, supraorbital
canal.

DISCUSSION Stage III) can be distinguished histologically. If only

The development of the cranial lateral line canals
in Archocentrus nigrofasciatus is described as a se-
ries of four stages. Neuromasts are apparent in the
epithelium before the beginning of canal morpho-
genesis. Neuromast differentiation was not observed
because it occurs in individuals ⬍5.5 mm SL. The
formation of bony canal walls, canal enclosure, and
canal ossification (the intramembranous ossification
of the canal roof) are sequential and discrete stages
in the development of lateral line canals. These
stages are difficult to distinguish without histologi-
cal analysis and they are not even distinguished in
some histological studies (Kapoor, 1970). Canal en-
closure and appearance of canal pores (Stage II vs.
SEM is used for analysis, canal enclosure and ossi-
fication (Stage III vs. IV) can only be inferred (e.g.,
Webb, 1989c). In most cases, cleared and stained
material can only reveal the timing of the appear-
ance (ossification) of ossified canal walls (Stage II)
and ossified canal roof (Stage IV) (e.g., Cubbage and
Mabee, 1996), but cannot determine when canal en-
closure (Stage III) occurs. This underscores the im-
portance of using both histology and SEM for such
an analysis.
The fact that one neuromast is located between
adjacent canal pores in the supraorbital and man-
dibular canals. This provides additional evidence
that this pattern of canal neuromast distribution
 CRANIAL LATERAL LINE CANAL DEVELOPMENT
Fig. 7. Late stages of morphogenesis of the portion of the
supraorbital (SO) canal in the frontal bone of Archocentrus nigro-
fasciatus. A: Canal is enclosed (Stage III) and ossified canal walls
(arrowheads, pink) extend around canal neuromast (6.5 mm SL,
Stage III) Scale bar ⫽ 50␮m. B: Low-power micrograph showing
the retina and brain showing the SO canal segment with an
ossified roof (11.5 mm SL, Stage IV) Scale bar ⫽ 200 ␮m. C:
Closeup of B. Canal is enclosed around neuromast with ossified
roof (arrows, pink). Scale bar ⫽ 50␮m. b, brain; e, retina; n,
supraorbital canal neuromast; fr, frontal bone; n, canal neuro-
mast; so, supraorbital canal.
and canal development is characteristic of acti-
nopterygian fishes (Webb and Northcutt, 1997). Ca-
nal development is initiated in the vicinity of indi-
vidual neuromasts, an observation that is consistent
53
with other reports in the literature (Amia, Allis,
1889; Amia, Pehrson, 1922; Polypterus, Pehrson,
1958; Ophicephalus, Kapoor, 1970). This supports
the hypothesis that the distribution of neuromasts
within the cranial canals of adult teleost fishes can
be predicted by the pattern of canal development
(Webb, 1989a; Webb and Northcutt, 1997).
In Archocentrus nigrofasciatus, the development
of individual segments of the SO and MD canals
occurs asynchronously. Histological data show that
canal segments appear to develop sequentially
within a given bone (e.g., frontal), but that enclosure
and ossification do not occur sequentially across
bone articulations (Table 1; see also Kapoor, 1970;
Jollie, 1984a; Mabee, 1993). This suggests that the
development of the canals in different bones occurs
somewhat independently of one another. For in-
stance, in the SO canal, canal segments appear to
enclose and ossify in a caudal to rostral direction
and the canal segment in the nasal bone is the last
to develop. In the MD canal, canal development ap-
pears to be caudal-rostral in the dentary; develop-
ment of the canal segment in the anguloarticular
occurs later. Asynchronous development of canal
segments within a canal (Lekander, 1949; Kapoor,
1961) and variation in the order to development of
canal segments and the overall timing of the devel-
opment of different canals have been noted in sev-
eral taxa (Allis, 1889; Lekander, 1949; Kapoor,
1961). Thus, it is suggested that variation in the
order and timing of the development of individual
canal segments, and the degree to which the canal
segments ossify and adjacent segments fuse (result-
ing in a single pore between adjacent canal seg-
ments), could serve to explain variation in the adult
morphology of the lateral line canal system among
teleosts (Webb, 1989a,b).
Functional Implications of the Pattern and
Timing of Canal Development
The pattern of lateral line canal development and
growth in bony fishes described here also has impor-
tant functional consequences. For instance, we have
shown that the SO and MD canals enclose (Stage
III) at a minimum diameter of about 50 ␮m, and
that the duration of Stage III (enclosed, without
ossified roof) is longer on average than the preceding
stage (Stage II, groove) in the SO canal (Tables 1, 2).
The hydrodynamic environment (e.g., a function of
the thickness of the boundary layer and the magni-
tude of the Reynolds number) in the vicinity of a
neuromast on a flat epithelium, in a groove, in a
pored canal with an epithelial roof, and in a pored
canal with a bony roof may be very different. Thus,
the diameter of the canal at enclosure, the duration
of each stage, and the timing of the development of
the entire canal system will have important impli-
cations for lateral line function in an ontogenetic
context. Further, the growth of neuromasts and the
54 M.L. TARBY AND J.F. WEBB

Fig. 8. Morphogenesis of
the mandibular (MD) canal in
Archocentrus nigrofasciatus.
A: Stage I: presumptive canal
neuromast (n) in the epithe-
lium overlie bone ossification
(arrow) surrounding the Meck-
el’s cartilage (m) (6.0 mm SL).
B: Stage II: canal neuromast in
groove (pink) (7.0 mm SL). C:
Stage III: the canal segment is
enclosed but the canal roof is
not ossified. Arrows indicate
distal ends of ossified canal
walls and dentary bone (de)
deep to neuromast (8.0 mm
SL), D: Stage IV: the canal is
enclosed around neuromast
and the canal roof is com-
pletely ossified (pink). The an-
guloarticular bone articulates
with the dentary and sur-
rounds the Meckel’s cartilage
(11.5 mm SL). Scale bars ⫽ 50
␮m. ar, anguloarticular bone;
de, dentary bone; m, Meckel’s
cartilage; n, canal neuromast.

increase in canal diameter after canal enclosure ap-
pear to be correlated (Fig. 10). Thus, it follows that
both of these processes must be coordinated with the
growth of the dermal bone in which the canal is
incorporated. In this way, the functional properties
of both the lateral line canal system, correlated with
major differences in canal morphology (e.g., narrow
vs. widened canals; Denton and Gray, 1988, 1989),
and the dermal bones, which play essential roles in
feeding and respiration (e.g., preopercular and man-
dible) or in the protection of the central nervous
system (e.g., frontal, nasal), are preserved through
ontogeny.
Growth of Lateral Line Canals
Canal diameter in the vicinity of individual neu-
romasts increases linearly after initial canal enclo-
sure (Stages III, IV), and canal diameter may double
as a fish increases in size by only a few millimeters
(Fig. 9). Thus, even at early stages of canal develop-
ment the bony canal walls must continuously accom-
modate the increasing size of the canal lumen by
adjusting their growth trajectories. Given that canal
diameter continues to increase after initial ossifica-
tion (when fishes are less than 10 mm in length),
dramatic changes in the dimensions of the dermal
bones, lateral line canals, and canal neuromasts are
expected to occur in concert with fish growth.
Changes in the size of flat dermal cranial bones can
easily occur as the result of changes in osteogenic
and osteoclastic fronts along the edges of dermal
bone matrix, but an increase in the diameter of the
lumen of the lateral line canals requires a different
process. Merillees and Crossman (1973) suggested
that remnants of the placodal precursors of neuro-
masts form a “connecting strand” (Clapp, 1889) com-
posed of cells that are responsible for the excavation
of the inside of the canal, allowing its diameter to
increase. Recently, it has been suggested that the
lumen of a tubular canal (the “praeopercular shaft,”
Witten and Villwock, 1997) increases in diameter
due to the activity of osteoclasts.
Do Lateral Line Bones in Archocentrus
nigrofasciatus Develop from One or Two
Bony Components?
The results of this study show that both the SO
and MD canals become incorporated into dermal
bones, as parallel canal walls extend upward from
the underlying dermal bone and fuse over individual
neuromasts to form individual ossified canal seg-
ments. Thus, the pattern of canal development ob-
served in the SO and MD canals of Archocentrus
nigrofasciatus does not appear to follow a two-
component pattern of development (e.g., Lekander,
1949) because two distinct bony elements (a tubular
laterosensory element and a flat underlying lamel-
lar element) could not be distinguished at early
stages of lateral line bone development. In addition,
the pattern of development described by Sire and
Huyseunne (1993), in which only a canal roof (but
not a tubular canal) appears to form separately from
the underlying lamellar component before fusing
with it, is not seen in A. nigrofasciatus. Thus, the
observed pattern of development in the SO and MD
canals in A. nigrofasciatus appears to demonstrate a
one-component pattern. It is possible that a two-
component pattern is present (where the laterosen-
 CRANIAL LATERAL LINE CANAL DEVELOPMENT
Fig. 9. Ontogenetic increases in the mean (left and right) diam-
eter of enclosed (Stage III) and ossified (Stage IV) canal segments in
the vicinity of individual canal neuromasts in the (A) supraorbital
and (B) mandibular canals of Archocentrus nigrofasciatus. A: Linear
regressions for the diameter of each of the five SO canal segments
were statistically significant (P ⬍ 0.01, geometric mean functional
regression [Ricker, 1973]). The r2 ⫽ 0.88, 0.93, 0.74, 0.75, 0.76, for
SO1–5, respectively. B: Canal segments in the vicinity of MD1, 2,
and 4 enclosed relatively late and no statistically significant trends
were noted, but a linear trend for MD3 was statistically significant
(P ⬍ 0.005; r2 ⫽ 0.76, geometric mean functional regression [Ricker,
1973]). See Figure 2 for locations of neuromasts in canals.
sory component is closely situated to the underlying
lamellar component), but could not be resolved using
the histological methodology employed (Kapoor,
1970). In this case, higher-resolution microscopic
55
Fig. 10. Relationship of mean (left and right) canal diameter
and mean neuromast width in the (A) supraorbital and (B) man-
dibular canals of Archocentrus nigrofasciatus after canal enclo-
sure (Stage III, IV). A: Linear regressions for each of the five SO
canal segments were statistically significant (P ⬍ 0.01). The r2 ⫽
0.96, 0.70, 0.54, 0.95, 0.95, for neuromasts SO1–5, respectively,
show a high correlation of neuromast width and canal diameter.
B: The MD canal encloses later than the SO canal. The relation-
ship of canal diameter and neuromast width showed a significant
linear trend only in the vicinity of neuromast MD3 (P ⬍ 0.01, r2 ⫽
0.77) but, the r2 values were still high for all canal segments (r2 ⫽
0.98, 0.70, 0.77, 0.88, for MD1– 4, respectively).
analysis (e.g., TEM) and/or a cell-level analysis of
the process of ossification (e.g., histochemical or mo-
lecular assays for osteoblast and osteoclast activity,
e.g., Witten, 1997; Witten and Villwock, 1997) could
be employed in order to distinguish between a one-
56 M.L. TARBY AND J.F. WEBB
component and a two-component pattern of lateral
line canal development. However, it should be noted
that these older histological studies clearly demon-
strate the presence of two bony components in sev-
eral taxa using histological methods similar to those
used in this study. Thus, it is concluded that the SO
and MD canals in A. nigrofasciatus develop follow-
ing a one-component pattern. The pattern of devel-
opment of the other cranial bones bearing lateral
line canals in this species awaits further analysis.
A two-component pattern of development has been
described in different canals in a number of teleost
fishes, but the description of a one-component pattern
of development like that described here in Archocen-
trus nigrofasciatus is available for only a few species.
Interestingly, reports of two-component lateral line
bones in the literature appear not to be uniformly
distributed among fishes. The older histological stud-
ies document the development of lateral line bones
from two bony components in non–teleost actinoptery-
gians (Amia, Allis, 1889; Pehrson, 1922, 1940; Lepisos-
teus, Aumonier, 1941; deBeer, 1985), elopomorph te-
leosts (Muraenidae, Allis, 1903; Anguilla, deBeer,
1985), basal euteleosts, including esocids (Esox, Pehr-
son, 1944), ostariophysans (Kindred, 1919; Lekander,
1949; Kapoor, 1961; Reno, 1966), and salmonids (re-
viewed in Pehrson, 1922; Jollie, 1984b; deBeer, 1985).
In contrast, Kapoor (1961) described “lamellar exten-
sions developing from the membranous component of
the frontal” (a one-component pattern) in Channa (⫽
Ophicephalus), a derived teleost. Of the 10 species of
derived teleosts described by deBeer (1985), develop-
ment of the lateral line canals is only mentioned with
reference to Sygnathus and Cyclopterus, two highly
modified taxa, so patterns and trends cannot be dis-
cerned. Branson and Moore (1962) described the lat-
eral line bones in centrarchids (Order Perciformes) as
“composite,” citing work by Stensio (1947) who de-
scribed “two component bones” in other taxa, but they
did not provide any ontogenetic data to back up this
claim. In contrast, Mabee (1993) describes what ap-
pears to be a one-component pattern of lateral line
bone development in centrarchids. Kapoor (1970) of-
fers an explanation of this apparent variation in pat-
tern of development among taxa suggesting that der-
mal bones have a “tendency to ’retreat’ deeper into the
softer tissues and away from the epithelium” in more
advanced teleosts. This would result in a closer asso-
ciation of the laterosensory and lamellar components
of a lateral line bone, which could explain variation in
ontogenetic patterns among different teleost taxa.
Thus, the pattern of development of the lateral line
canals in the dermal cranial bones may vary phyloge-
netically, with a transition from a two-component pat-
tern in basal teleosts to a one-component pattern in
more advanced taxa. Further histological analyses of
the development of lateral line canals in homologous,
lateral line canal-bearing bones in representative te-
leost taxa are needed in order to further evaluate this
hypothesis.
Summary
An accurate view of how lateral line canals be-
come integrated into the dermal bones of the skull of
bony fishes will contribute to our understanding of
several fundamental issues in the comparative mor-
phology and development of the fish skull. These
include 1) the developmental basis for variation in
the association of a lateral line canal with different
dermal cranial bones (Graham-Smith, 1978), 2) the
patterns and mechanisms of phylogenetic loss of
lateral line canals and/or dermal cranial bones in
fishes and tetrapods (Kapoor, 1970; deBeer, 1985),
and 3) patterns of phylogenetic variation in the mor-
phology of the lateral line canal system among fishes
(see Webb, 1989b for review). Such data could pro-
vide a context for a reevaluation of homologies
among dermatocranial bones (see Allis, 1889; re-
viewed by Kapoor, 1970). Finally, an appreciation of
the pattern of development of lateral line canals and
of the dermal bones with which they are associated
can provide an ontogenetic context for the testing of
meaningful hypotheses concerning the developmen-
tal mechanisms that underlie this association. The
putative inductive relationship between neuromasts
and dermal cranial bones has been discussed by
several workers (e.g., Branson and Moore, 1962;
Reno, 1966; Kapoor, 1970; Patterson, 1977; Schaef-
fer, 1977; Graham-Smith, 1978; Northcutt and
Gans, 1983) and the mechanism underlying this as-
sociation has been stated to be a major problem by
deBeer (1985) and Hall and Hanken (1985), but it
remains unexplored, and thus unresolved.
ACKNOWLEDGMENTS
We thank Dr. Tom Miyake (Dalhousie University)
for helpful suggestions for histological protocols, Dr.
Norman Dollahon, who provided assistance with
preparation of SEM material, and Dr. Michael Rus-
sell, who provided statistical expertise. Jeff DeSalvo,
Joanna Dyer, and Dhira Khosla assisted in the rear-
ing and sampling of fishes. Leo Smith and Sean
Maher assisted with the preparation of the figures.
This work is in partial fulfillment of requirements
for the M.S. degree at Villanova University to MLT.
LITERATURE CITED
Adriaens D, Varraes W, Taverne L. 1997. The cranial lateral-line
system in Clarias gariepinus (Burchell, 1822) (Siluroidei: Clari-
idae): morphology and development of canal related bones. Eur
J Morphol 35:181–208.
Allis EP. 1889. The anatomy and development of the lateral line
system in Amia calva. J Morphol 2:463–566.
Allis EP. 1903. The lateral sensory system in the Muraenidae. Int
Mschr Anat Physiol 20:125–170.
Aumonier JF. 1941. Development of the dermal bones in the skull
of Lepidosteus osseus. Q J Microsc Sci 83:1–33.
Axelrod HR, Vorderwinkler W. 1980. Encyclopedia of tropical
fishes. Neptune, NJ: T.F.H.
 CRANIAL LATERAL LINE CANAL DEVELOPMENT
Branson BA, Moore GA. 1962. The lateralis components of the
acoustico-lateralis system in the sunfish family Centrarchidae.
Copela 1962:1–108.
Clapp CM. 1889. The lateral line system of Batrachus tau. J
Morphol 15:223–264.
Coombs S, Janssen J, Webb JF. 1988. Diversity of lateral line
systems: Evolutionary and functional considerations. In: Atema
J, Fay RR, Popper AN, Tavolga WN, editors. Sensory biology of
aquatic organisms. New York: Springer-Verlag. p 553–593.
Cubbage CC, Mabee PM. 1996. Development of the cranium and
paired fins in the zebrafish Danio rerio (Ostariophysi, Cyprin-
idae). J Morphol 229:121–160.
de Beer GR. 1985. The development of the vertebrate skull. Chi-
cago: University of Chicago Press.
Denton EJ, Gray JAB. 1988. Mechanical factors in the excitation
of the lateral lines of fish. In: Atema J, Fay RR, Popper AN,
Tavolga WN, editors. Sensory biology of aquatic animals. New
York: Springer-Verlag. p 595– 618.
Denton EJ, Gray JAB. 1989. Some observations on the forces acting
on neuromasts in fish lateral line canals. In: Coombs S, Gorner P,
Munz H, editors. The mechanosensory lateral line: neurobiology
and evolution. New York: Springer-Verlag. p 229–246.
Devillers C. 1947. Recherches sur la crâne dermique des téléos-
téens. Ann Paléontol 33:1–94.
Graham-Smith W. 1978. On the lateral lines and dermal bones in
the parietal region of some crossopterygian and dipnoan fishes.
Philos Trans R Soc Lond B 282:41–105.
Hall BK. 1986. The role of movement and tissue interactions in
the development and growth of bone and secondary cartilage in
the clavicle of the embryonic chick. J Embryol Exp Morphol
93:133–152.
Hall BK, Hanken J. 1985. Foreword. In: de Beer GR. The devel-
opment of the vertebrate skull. Chicago: University of Chicago
Press. p vii–xxviii.
Hanken J, Thorogood P. 1993. Evolution and development of the
skull: the role of pattern formation. Trends Ecol Evol 8:9 –14.
Jollie M. 1984a. Development of cranial and pectoral girdle bones
of Lepisosteus with a note on scales. Copeia 1984:476 –502.
Jollie M. 1984b. Development of the head skeleton and pectoral
girdle of salmons, with a note on the scales. Can J Zool 62:1757–
1778.
Kapoor AS. 1961. The time and order of formation of sensory
canals in the fishes Ophicephalus punctatus (Ophicephalidae)
and Wallago attu (Siluridae). Copeia 1961:176 –181.
Kapoor AS. 1970. Development of dermal bones related to sensory
canals of the head in the fishes Ophicephalus punctatus Bloch
(Ophicephalidae) and Wallago attu Bl. & Schn. (Siluridae). Zool
J Linn Soc 49:69 –97.
Kapoor AS. 1972. Development of latero-sensory canal bones in
the fish Ophicephalus punctatus. Copeia 1962: 854 – 855.
Kindred JE. 1919. The skull of Amiurus. Ill Biol Monogr 5:120.
Lekander B. 1949. The sensory line system and the canal bones in
the head of some Ostariophysi. Acta Zool (Stockh) 30:1–131.
Mabee PM. 1993. Phylogenetic interpretation of ontogenetic
change: sorting out the actual and artefactual in an empirical
case study of centrarchid fishes. Zool J Linn Soc 107:175–291.
Mabee PM, Trendler TA. 1996. Development of the cranium and
paired fins in Betta splendens (Teleostei: Percomorpha): in-
traspecific variation and interspecific comparisons. J Morphol
227:249 –287.
Merillees MJ, Crossman EJ. 1973. Surface pits in the family esoci-
dae. II. epidermal dermal interaction and evidence for aplasia of
the lateral line sensory system. J Morphol 141:321–344.
Morris SL, Gaudin AJ. 1982. Osteocranial development in the
viviparous surfperch, Amphistichus argenteus (Pisces: Embioto-
cidae). J Morphol 174:95–120.
Münz H. 1986. What influences the development of canal-and su-
perficial neuromasts? Ann Mus R Afr Centr Sc Zool 251:85– 89.
Münz H. 1989. Functional organization of the lateral line periph-
ery. In: Coombs S, Gorner P, Münz H, editors. The mechanosen-
57
sory lateral line: neurobiology and evolution. New York:
Springer-Verlag. p 285–297.
Neave DA. 1986. The development of the lateral line system in
plaice (Pleuronectes platessa) and turbot (Scophthalmus maxi-
mus). J Mar Biol Assoc UK 66:683– 693.
Northcutt RG, Gans C. 1983. The genesis of neural crest and
epidermal placodes: a reinterpretation of vertebrate origins. Q
Rev Biol 58:1–58.
Patterson C. 1977. Cartilage bones, dermal bones and membrane
bones, or the exoskeleton versus the endoskeleton. In: Andrews
SM, Miles RS, Panchen AL, editors. Problems in vertebrate
evolution. New York: Academic Press. p 77–121.
Pehrson T. 1922. Some points in the cranial development of
teleostean fishes. Acta Zool (Stockh) 3:1– 63.
Pehrson T. 1940. The development of dermal bones in the skull of
Amia calva. Acta Zool 211:1–50.
Pehrson T. 1944. The development of latero-sensory canal bones
in the skull of Esox lucius. Acta Zool (Stockh) 25:135–157.
Pehrson T. 1958. The early ontogeny of the sensory lines and the
dermal skull in Polypterus. Acta Zool 39:241–258.
Poling KR, Fuiman LA. 1997. Sensory development and concur-
rent behavioural changes in Atlantic croaker larvae. J Fish Biol
51:402– 421.
Pothoff T. 1983. Clearing and staining technique. In: Ontogeny
and systematics of fishes. Special publication no. 1. American
Society of Ichthyologists and Herpetologists.
Reno HW. 1966. The infraorbital canal, its lateral-line ossicles
and neuromasts, in the minnows Notropis volucellus and N.
buchanani. Copeia 1966:403– 413.
Ricker WE. 1973. Linear regressions in fishery research. J Fish
Res Bd Can 30:409 – 434.
Schaeffer B. 1977. The dermal skeleton in fishes. In Andrews SM,
Miles RS, Panchen AL, editors. Problems in vertebrate evolu-
tion. New York: Academic Press. p 25–52.
Sire J-Y, Huysseune A. 1993. Fine structure of the developing
frontal bones and scales of the cranial vault in cichlid fish
Hemichromis bimaculatus (Teleostei, Perciformes). Cell Tissue
Res 273:511–524.
Stensio EA. 1947. The sensory lines and dermal bones of the
cheek in fishes and amphibians. Kungl Svenska Vetens Handl
24:1–195.
Swertzoff AN. 1926. Studies on the bony skull of fishes. I. Struc-
ture and development of the bony skull of Acipencer ruthenus. Q
J Microsc Sci 70:451– 640.
Tarby ML. 1998. Morphogenesis of the lateral line canals in the
cranial dermal bones of the cichlid, Archocentrus nigrofascia-
tus. MS Thesis, Villanova University, Villanova, PA.
Wagemans F, Focant B, Vandwalle P. 1998. Early development of
the cephalic skeleton in the turbot. J Fish Biol 52:166 –204.
Webb JF. 1989a. Developmental constraints and the evolution of
the mechanosensory lateral line system in teleost fishes. In:
Coombs S, Görner P, Münz H, editors. The mechanosensory
lateral line: neurobiology and evolution. New York: Springer-
Verlag. p 79 – 89.
Webb JF. 1989b. Gross morphology and evolution of the mech-
anoreceptive lateral-line system in teleost fishes. Brain Behav
Evol 33:205–222.
Webb JF. 1989c. Neuromast morphology and lateral line trunk
ontogeny in two species of cichlids: an SEM study. J Morphol
202:53– 68.
Webb JF, Northcutt RG. 1997. Morphology and distribution of pit
organs and canal neuromasts in non-teleost bony fishes. Brain
Behav Evol 50:139 –151.
Witten PE. 1997. Enzyme histochemical characteristics of osteo-
blasts and mononucleated osteoclasts in a teleost fish with
acellular bone (Oreochromis niloticus, Cichlidae). Cell Tissue
Res 287:591–599.
Witten PE, Villwock W. 1997. Growth requires bone resorption at
panguloarticular skeletal elements in a teleost fish with acel-
lular bone (Oreochromis niloticus, Teleostei, Cichlidae). J Appl
Ichthyol 13:149 –158.