Organization of the ectodermal nervous structures in medusae: cubomedusae.
Cubomedusae are fast swimmers that are capable of rapid, accurate turning, up to 180[degrees] in as few as two swim contractions (Garm et al., 2007a). Generally speaking, box jellyfish are neritic species inhabiting a variety of environments, including mangrove roots, kelp forests, and coral reefs. Many cubomedusae actively reside in these habitats rather than being passively distributed by currents and waves like many other jellyfish species. Such active residence requires swimming agility supported by excellent sensory systems (Coates, 2003). This raises critical questions about how cubomedusae produce such accurate swimming adjustments, and in a more general sense, how these radially symmetric, gelatinous animals respond to asymmetric stimuli from the environment to drive their acute and long-term behavioral responses.
Complex swimming behaviors necessitate some form of central processing to sort a milieu of sensory information into appropriate motor outputs. Garm et al. (2007b) define the cubozoan central nervous system (CNS) as containing the integrative elements of the rhopalia and the nerve ring. The cubomedusan nervous system thus shows both compression of neuronal networks into the tract-like nerve ring, and co-localization of sensory structures and ganglion-like neuronal accumulations in the rhopalia. In comparison, scyphomedusae possess rhopalia but without the connecting tracts of a nerve ring, while hydromedusae have a pair of complex nerve rings representing their centralized nervous systems (see Satterlie, 2011). Neural condensation into the nerve rings may be associated with a more diverse behavioral repertoire in the Hydrozoa and Cubozoa, allowing more rapid and accurate behavioral adjustments. The compression of multiple conducting systems into a nerve ring, or rings, along with the accumulation of neurons and sensory cells into rhopalia, could be considered the cnidarian equivalent of centralization in bilaterians (Satterlie, 2002, 2011).
Despite the existence of a centralized nervous system in some cnidarians, most retain conducting components in the form of nerve nets (Satterlie, 2011). Immunohistochemical staining of scyphomedusae with monoclonal antibodies against [alpha]- or [beta]-tubulin shows a subumbrellar nerve net that has the distribution and neuronal characteristics of the motor nerve net that controls the swim musculature (Giant Fiber Nerve Net, also called the Motor Nerve Net; Horridge, 1956a, b; Schwab and Anderson, 1980; Anderson and Schwab, 1981). In addition, antibodies against RFamides, specifically the neuropeptide FMRFamide, stain a second, distinct subumbrellar nerve net that fits the anatomical and physiological properties of the Diffuse Nerve Net in scyphomedusae (Grimmelikhuijzen, 1985; Anderson et al., 1992; Grimmelikhuijzen et al., 1996, 2002; Satterlie 2002). Other parts of scyphomedusae, including tentacles, manubrium, and exumbrella, show similar double staining, with the FMRFamide-immunoreactive networks including putative sensory cells.
In cubomedusae, the FMRFamidergic system in the subumbrella is almost exclusively contained within the rhopalia and nerve ring, without the diffuse nerve net organization seen in scyphomedusae (Satterlie, 2002, 2011). On the other hand, tubulin-immunoreactive nerve networks are found throughout the subumbrellar areas containing swim musculature, including the subumbrella proper, the velarium, and the velarial frenula. Comparison of the scyphomedusan and cubomedusan nervous systems is important since the cubomedusae were originally included within the class Scyphozoa, and only recently were elevated to class status (class Cubozoa; Werner et al., 1971, 1976). From a behavioral and neurobiological perspective, this association goes back to the work of Romanes (1885) who provided clear distinctions in the nervous organization and behavior of the covered-eyed medusae (Scyphozoa) and the naked-eyed medusae (Hydrozoa). The initial association of scyphomedusae and cubomedusae was supported by similar basic characteristics of their swimming systems, where the cubomedusae clearly fit within the category of Romanes' covered-eyed medusae (Passano, 1965; Satterlie, 2002). This warrants a thorough comparison of the neuronal architecture of representatives of the two groups, particularly in light of their very different swimming agility and their differing class status, and in light of unique characteristics of the Cubozoa that include the presence of lens-bearing rhopalial eyes and a nerve ring that interconnects the rhopalia.
Within the swim system of cubomedusae, the density of neurons in the velarial nerve net appears to exceed that of the subumbrellar net (Satterlie, 2002). This may have implications for turning behavior since asymmetric contractions of the velarium underlie nozzle formation of the bell aperture for directional swimming (Gladfelter, 1973; Petie et al., 2011). The motor nerve net of the perradial frenula is also notable in that neuronal density appears greater than that of the velarium or subumbrella (based on qualitative assessments; Satterlie et al., 2005). The frenular network is continuous with the velarial network, suggesting that these two muscle sheets contract together (Satterlie et al., 2005). Should network density impact physiological characteristics of the network, such as conduction velocity and facilitation rate, the frenula might play a key role in shaping the velarium for directional fluid ejection and turning, much like exumbrellar radial muscle in the velum of hydromedusae (Gladfelter, 1973; Satterlie et al., 2005).
Here we provide a quantitative confirmation of network density differences in the subumbrella, velarium, and frenula within and between cubomedusan species, as part of a thorough examination of their ectodermal nervous system. This includes a further test of a prediction formed from our work on scyphomedusae: staining with anti-tubulin antibodies labels networks associated with swim muscle sheets in cubomedusae, whereas the FMRFamide antibody labels networks that include putative sensory components. The results of the overall immunohistochemical work are couched in terms of the unique behavioral responses of cubomedusae, including feeding and protective crumpling.
Materials and Methods
Specimens of Carybdea marsupialis Linnaeus, 1758 were collected by scuba divers off the coast of Santa Barbara, California, and shipped overnight to the Center for Marine Science at the University of North Carolina Wilmington (UNCW). Specimens were housed in a chilled, recirculating aquarium at ~15[degrees]C and a narrow salinity range of 31-33 ppt. Specimens of Tripedalia cystophora Conant, 1897 were collected from La Isla Magueyes in Puerto Rico by Dr. Alina Szmant (UNCW) and her graduate students and shipped overnight in 1-liter plastic containers. Jellyfish were placed in fixative immediately upon arrival. Otherwise they were maintained in a recirculating aquarium at a salinity of 31-33 ppt and temperature of 20-22[degrees]C very briefly (1-2 days) until needed. Tamoya haplonema Muller, 1859 and Chiropsalmus quadrumanus Muller, 1859 were collected locally by bottom trawls in 6 to 15 m of water. Specimens were dissected and fixed immediately on return to the laboratory. These large medusae were used for corroboration of results from the two smaller species, and were not subjected to the quantitative tests of neurite density and neurite orientation.
Prior to dissection, animals were relaxed in a 1:2 mixture of isotonic Mg[Cl.sub.2] (0.33 mol [l.sup.-1]) in seawater for at least 5 min. The preparations were rinsed with seawater once quickly prior to immersion in fixative to minimize precipitate formation. Excised tissues were fixed for 4 h in 4% paraformaldehyde in 0.1 mol [l.sup.-l] phosphate buffer (pH 7.4) at room temperature (20-22[degrees]C). Fixed material was washed at least four times in 0.1 mol [l.sup.-1] phosphate buffer (pH 7.4) containing 0.01% Triton X-100 or 0.05% Tween 20. All washes were 15 min in length unless noted otherwise. The tissue was dehydrated and then rehydrated through a graded ethanol series (50%, 70%, 90%, 100%, 100%, 90%, 70%, 50%) to further permeabilize the samples. After blocking in 5% goat serum (Sigma Chemical G6767) in phosphate buffer for 2-4 h, the tissue was incubated for 24-48 h in primary antibody (anti-[alpha]-tubulin, mouse, Developmental Studies Hybridoma Bank 12G10; anti-[beta] tubulin, mouse, Developmental Studies Hybridoma Bank AA12.1; anti-FMRFamide, rabbit, Millipore AB15348; anti-actin, rabbit, Sigma A5060), washed in phosphate buffer (four to eight changes), and incubated for 12-24 h in fluorescent-tagged secondary antibody (Sigma: Goat anti-mouse [FITC] F5262; Goat anti-mouse [TRITC] T7782; Goat anti-rabbit [FITC] F9887; Goat anti-rabbit [TRITC] T6778). After another wash series in phosphate buffer, tissue pieces were cleared and mounted in a 9:1 mixture of glycerol and 50 mmol [l.sup.-1] Tris buffer (pH 9). Control preparations were included in parallel where tissue was incubated in fluorescently tagged secondary antibody, but in the absence of either primary antibody. No labeling of neurites was observed in any controls. Additionally, no significant co-localization of staining was observed for tissue incubated with both the anti-FMRFamide and antitubulin antibodies (see Satterlie and Eichinger, 2014, for a discussion of cross reactivity with these antibodies).
Preparations were examined using a Nikon epifluorescence microscope with FITC or TRITC filter cubes, and photographed with a SPOT camera (Diagnostic Instruments, Inc., Sterling Heights, MI). Alternatively, preparations were examined using an Olympus FV1000 laser scanning confocal microscope (Olympus, Center Valley, PA). No variations were found with the stated ranges of fixation and incubations, so they were adjusted for experimental convenience.
Using the confocal microscope, optical slices of 0.5-1-[micro]m thickness were collected from two channels: one for the 488-nm (FITC) signal and one for the 543-nm (TRITC) signal. Twenty to sixty images were collected for each z-stack, depending on the tissue viewed, and nerve net reconstructions were visualized using Olympus Fluoview ver. 1.6a software. The perradial frenula are three-dimensional muscle sheets, oriented perpendicular to the 2-dimensional subumbrella and velarium. In Carybdea, this muscle sheet was excised and spread flat for viewing. In Tripedalia, the frenula were too small for excision, but specimens were compressed tightly between the slide and cover slip to flatten them as much as possible. Extreme care was taken to include only a single plane of the frenular muscle sheet for inclusion in neuron counts. No differences in tissue depth were noted between specimens, and any differences in zstack depth were due to very small folds in the tissue.
Compressed z-stack images of a-tubulin-labeled nerve nets from the swim musculature were printed onto 9 X 11-in paper, and 100 X 100-[micro]m quadrants were drawn using a stencil on each image to cover as much area as possible. These measures were taken to eliminate regions of reduced fluorescence. All regions of each z-stack image were sampled, except regions of reduced fluorescence due to minor folds in the tissue that could not be included without also including tissue layers that might skew neurite counts. The number of neurites in each quadrant was manually counted, and values for the mean and standard error of the mean (SEM) were calculated for each z-stack image. In the case of the frenulum, neurite counts were administered only to flattened portions of this 3-dimensional muscle sheet in Tripedalia and Carbydea.
Each image was included as a sample for statistical analysis, but no more than two images originated from any single specimen. Data were analyzed using JMP 7.0.7 (SAS Institute, Cary, NC). One-way ANOVAs and a post hoc Tukey HSD test were used to test for within-species differences in network density between muscle sheets. A two-way ANOVA was used to test for group effects between species and muscle sheets. A post hoc Tukey HSD test was used to detect differences between groups, and Student's t-tests were used to make pairwise comparisons of means to further clarify results. P < 0.05 was considered significant in all tests.
The angular orientation of each neurite was found relative to the axis of the bell margin designated as 180[degrees], and where 90[degrees] would indicate a radial orientation. A small protractor was used to measure the angle of displacement of individual neurites from the axis of the bell margin. A line was superimposed onto particularly sinuous neurites to determine the overall vector of displacement from the bell margin. Angle measures were sorted into eighteen 10[degrees] bins, spanning the range of 0 to 180[degrees]. Chi-square tests were utilized for each species to test for significant differences in neurite orientation (number of neurites per bin) within and between muscle sheets. P < 0.05 was considered significant for all tests.
Compressed z-stack images of a-tubulin-labeled nerve nets from the swim musculature of Tripedalia were uploaded into ImageProPlus (ver. 184.108.40.2066; Media Cybernetics, Bethesda, MD). The mean diameters of a random sampling of neurites were measured in subumbrellar, velarial, and frenular muscle sheets of four specimens of each species. Statistical analysis consisted of a one-way ANOVA, followed by a Tukey HSD test, administered through JMP 7.0.7 (SAS Institute, Cary, NC).
Transmission electron microscopy
Jellyfish tissue was fixed in 2.5% glutaraldehyde, 2% paraformaldehyde, and 3% sucrose in 0.15 mol [l.sup.-1] sodium cacodylate buffer at room temperature (20-22[degrees]C) and pH (7.4). After 5-6 days in fixative, tissue was washed in sodium cacodylate buffer at physiological pH (7.4), followed by four more washes at 20-25 min each. Tissue was post-fixed in 1% osmium tetroxide in (0.15 or 0.2 mol [1.sup.-1]) sodium cacodylate buffer for 1 h (longer if tissue pieces were large) at room temperature, rinsed for 15 min in 0.15 mol [l.sup.-1] sodium cacodylate buffer and 10 min in D1 water. Tissue was dehydrated in an ethanol series at 15-20-min per wash (50%, 70%, 95%, 100%, 100%) and transferred to pure acetone (wash for 15 min). Tissue was rotated for 1 h in a 1:1 acetone to SPURR epoxy resin mixture (Spurr, 1969), then rotated overnight in 100% SPURR. Fresh SPURR was made the next morning for embedding tissue, which was then placed in the oven until hardened. Samples were sectioned using an Ultracut-E ultramicrotome (Leica, Buffalo Grove, IL). Thin sections (~90 nm) were picked up onto Formvar-coated copper grids, and stained using uranyl acetate in 50% ethanol and Reynolds' lead citrate (Reynolds, 1963), for 20 min each. The specimen-containing grids were viewed on a Philips CM-12 transmission electron microscope (Philips Research, Briarcliff Manor, NY) operated at 80 kV and photographed using Kodak EM 4489 film. Negatives were developed and digitized using a Microtek Scanmaker 4 (Microtek Lab, Carson, CA).
The body plans of all four box jellyfish species utilized here are nearly identical with the exception of obvious differences in size and number of tentacles. The bell of adult Tripedalia cystophora is usually no more than 1-2 cm in height; it has three blade-like pedalia at each interradius and a single tentacle extending from each pedalium. This species is found in light shafts between mangrove roots, where it feeds on phototaxic copepods (Stewart, 1996). Carybdea marsupialis (formerly Carybdea rastonii) can grow to a bell height of 4 cm and has a single pedalium with a single tentacle at each interradius. This species is found in swarms over sandy bottoms in surge channels between kelp forests, where it feeds on mysids and larval fish (Matsumoto, 1995). Tamoya haplonema (up to 14 cm in bell height and 6 cm bell width) and Chiropsalmus quadrumanus (up to 9.5 cm in bell height and width) were found in sandy or rockybottom areas where they feed on shrimp and small fish.
Despite considerable differences in their ecology, nervous system design is virtually indistinguishable in these four species. Neurons in the rhopalia, nerve ring, and nerve nets of all species are immunoreactive to [alpha]- or [beta]-tubulin antibodies and the FMRFamide antibody.
The subumbrella, velarium, and frenulum are the three effector components of the swim system, and each muscle sheet has a unique motor network (Fig. 1). The subumbrellar muscle is the largest muscle sheet and lines the inner walls of the bell above the nerve ring, while the velarium is a continuation of the subumbrella below the nerve ring forming the the bell opening. Both of these muscle sheets contain circular, striated muscle fibers that, when activated, restrict the circumference of the bell cavity and the bell opening, respectively. The four perradial frenula are located below the nerve ring opposite the rhopalial niches. These frenula contain striated, radial muscle fibers that extend from its origin near the top of the rhopalial niche down to the bell opening.
Neurite morphology. Neurites of the subumbrella, velarium, and frenulum in Tripedalia appeared morphologically diverse. Fluorescent puncta extended along the length of velarial and subumbrellar neurites (Fig. 1). In addition, larger puncta occurred where neurites crossed one another. Puncta in the subumbrellar network were less prominent but were still found along the length of individual neurites. Puncta were not as apparent in the frenula, but the density of neurites may have obscured them.
In Tripedalia, each motor network possessed neurites as small as 1 [micro]m in diameter or as large as 5 [micro]m in diameter, although neurites of the latter size were less prevalent in the subumbrella and frenulum than in the velarium. A random sampling of neurites revealed that the subumbrella contained neurons with the smallest neurites (3.4 [+ or -] 0.16 jam) followed by the slightly larger frenular neurites (3.6 [+ or -] 0.15 [micro]m). Velarial neurite measurements (4.2 [+ or -] 0.25 [micro]m) were significantly larger in diameter than subumbrellar neurites, but not frenular neurites. Statistical analyses were conducted via a one-way ANOVA, and a post hoc Tukey HSD test (F(2,60) = 4.60, P = 0.0138).
The robust staining and stout appearance of velarial neurons in Carybdea, in comparison to those of the subumbrellar nerve net, prompted further microscopical analysis of neuron morphology. Under high magnification, some of the larger velarial neurites appeared to be bundles of smaller neurites, which did not show the characteristic punctuate appearance of individual neurites. Similar bundling was not found in any of the swim motor networks of Tripedalia, but it was found in all three networks of Carybdea. Neurite bundles represented a small percentage of the total neurite population and most often involved the tight apposition of two to three individual neurites (verified with electron microscopical examination--not shown). Due to bundling, absolute measurements of neurite diameters will be artificially high, and absolute measurements of neurite densities will be low. However, since all three motor networks in the swim system of Carybdea displayed bundling, relative densities in the three areas will likely show similar comparative proportions.
Neurite density. Initial evaluation of the motor nerve nets of the three muscle sheets indicated clear differences in the neuronal density, with the highest density in the frenula and the lowest in the subumbrella (Figs. 2, 3). To verify the qualitative observations, network density was analyzed statistically.
A one-way ANOVA was utilized to compare neurite densities between the muscle sheets in Tripedalia (n = 5). The three muscle sheets had significantly different neurite densities in the a-tubulin-IR motor nerve net (F(2,16) = 169.99, P < 0.001). Post hoc analysis via a Tukey HSD test revealed that frenular neurite density (22.83 [+ or -] 0.5) was significantly greater than that of the velarium (16.10 [+ or -] 0.37), which in turn exhibited a greater neurite density than the subumbrella (11.44 [+ or -] 0.31) (Fig. 4).
Carybdea (n = 3) exhibited a similar relationship using the same analysis (Fig. 2). Neurite densities in the separate muscle sheets were significantly different in the motor nerve net (F(2,12) = 25.57, P < 0.001). The frenulum possessed the greatest neurite density among the muscle sheets (20.37 [+ or -] 0.7), followed by the velarium (16.34 [+ or -] 0.4), and the subumbrella (11.28 [+ or -] 0.33), which possessed a significantly lower density than either of the other muscle sheets.
At the bell perradii, neurites from the nerve ring turned toward the perradial midline and condensed near the point of attachment of the frenulum apex to the subumbrellar side of the bell, opposite the rhopalial stalk (Fig. 3a). Neurites maintained this compressed network configuration throughout the length of the frenulum until they reached the bell margin (Fig. 3b, c). Some neurites splayed out into the velarium, but many turned back into the frenular nerve net before they reached this barrier, creating a zone of decreased neurite density between the two muscle sheets (Fig. 3b). A small population of neurites from the velarium turned 90[degrees] into the frenulum, but not in the middle of the frenulumvelarium junction.
Neurite orientation. The a-tubulin-IR swim motor nerve net is a major component of the peripheral nervous system in cubomedusae. In Tripedalia, neurites of the subumbrellar and velarial motor nerve nets exhibited no dominant directional trend (Fig. 4). Instead they conformed to the classic diffuse nerve net design where neurites were equally distributed at all angles relative to the axis of the bell margin (axis parallel to the bell opening, designated as 180[degrees]). Close to the nerve ring and the frenulum, velarial neurites were initially oriented more radially, but midway between the interradii and perradii, and well away from the nerve ring, the velarial network showed random neurite orientations. The frenulum exhibits a uniquely organized nerve net in which neurites align with the radial axis and radially arranged muscle fibers (Figs. 3, 4).
Comparison of the velarial and subumbrellar networks failed to detect any preferred orientation of neurites ([chi square](17, n = 423) = 13.07, P = 0.73). In comparisons of the frenulum with either the velarium or subumbrella, neurite angle measurements were significantly different between muscle sheets (velarium: [chi square](17, n = 369) = 120.14, P < 0.001; subumbrella: [chi square](17, n = 406) = 110.58, P < 0.001). Neurite angle measurements in the frenulum predominately conformed to the 61[degrees] to 80[degrees] bins (relative to the axis of the bell margin; Fig. 4f). These analyses indicate that neurite orientation in the frenulum is significantly distinct from that of the other muscle sheets.
Similar results were obtained for specimens of Carybdea (Fig. 4a, b, c). Neurites in the subumbrella and velarium were distributed at random angles relative to the axis of the bell margin. Specifically, they did not align with circular muscle fibers. In contrast, neurites of the frenular nerve net aligned with the radial muscle fibers, thus deviating from the classic diffuse nerve net organization (Figs. 3, 4c). Chi-square analysis indicated that frenular neurites exhibited different angular orientations in comparison to other muscle sheets in this species (velarium: [chi square](17, n = 217) = 119.40, P < 0.001; subumbrella: ^(17, n = 218) = 154.67, P < 0.001). Comparison of neurite angle measurements in the velarium and subumbrella were similar 0^(17, n = 259) = 28.83, P = 0.036). As in Tripedalia, neurite angle measurements in the frenulum of Carybdea showed a preferred orientation within the 61[degrees] to 80[degrees] bins, while neurite angle measurements in the subumbrella and velarium were randomly distributed. The cause for a slightly more platykurtic distribution of neurite angle measures in the frenula of Tripedalia relative to Carybdea was likely due to the much larger frenula in the latter that created a greater sampling area farther from the velarial nerve net.
Rhopalia and nerve ring
The rhopalia of cubomedusae exhibited such dense tubulin staining that individual neurites were impossible to distinguish. FMRFamide immunoreactivity in the rhopalium was localized to a small network of cells surrounding the attachment site of the rhopalial stalk, as previously described (Martin, 2004; Skogh et ah, 2006; Parkefelt and Ekstrom, 2009). This network had a consistent position and arrangement in all four species of cubomedusae, with a loose network of brightly fluorescent cells in the pacemaker region (Fig. 5). The network incompletely surrounded the rhopalial stalk in a horseshoe shape, with the open area in the anterior region (the side centered on the midline of the two complex eyes). In the posterior, superior region of the rhopalia, the network included a dense pad of immunoreactive epithelial cells, suggesting that a putative sensory pad is part of this network (Fig. 5). Surrounding the rhopalium, the epithelium of the rhopalial niche also contained a dense FMRFamide-IR network, much like that of the scyphozoans (Satterlie and Eichinger, 2014).
In agreement with Garm et al. (2007b) the [alpha]-tubulin antibody labeled two primary tracts and a more diffuse network in the rhopalial stalk. The two tracts connected to the nerve ring, which showed intense tubulin staining, much like the rhopalia (Fig. 6a). Neurons from the nerve ring diverged into both subumbrellar and velarial motor networks.
There were up to three FMRFamide-IR neuronal tracts in all sections of the nerve ring (Garm et al., 2007b; Satterlie, 2011; Fig. 6c). Unlike tubulin immunostaining, FMRF-amide-IR found few processes exiting the nerve ring, and they did not extend far into the subumbrella. Away from the nerve ring, no FMRFamide-IR neurons were found in either the subumbrella or velarium.
Radial muscle bands and manubrium
Just above the rhopalial niche in each bell quadrant, perradial bands of smooth muscle originated just below the nerve ring (Satterlie et al., 2005). Some [alpha]-tubulin-IR neurons from the nerve ring turned 90[degrees] into the smooth muscle bands, which appeared to be densely innervated compared to the surrounding subumbrella (Fig. 6a). When neurons from the subumbrellar nerve net made contact with the smooth muscle bands, they either crossed the band or converged in the radial direction with smooth muscle fibers. Outcroppings of the smooth muscle band turn 90[degrees] and extend for a short distance into the striated muscle sheet of the subumbrella (Satterlie et al., 2005). [alpha]-Tubulin-IR neurons did not follow or conform to these muscular outcroppings (Fig. 6b).
The smooth muscle bands and their respective networks extended upward through the subumbrella to the perradial ridges of the quadrangular manubrium that hangs from the roof of the bell. These bands formed muscular ridges that continued to the elongated lips of the manubrium, but became less defined and eventually diverged into circular muscle fibers. Flere, neurons from the smooth muscle band joined the manubrial nerve net.
Within the manubrium, neurons maintained a netlike organization that extended to the distal lips of the manubrium where neurons deviated from a random orientation and straightened into parallel tracts perpendicular to the manubrial edge (Fig. 7a). Ciliated cells that lined the edge of the manubrium stained very brightly with the anti-tubulin antibody.
At the perradii, FMRFamide-IR neuronal tracts in the nerve ring converged, and some bypassed the connection with the rhopalial stalk (Fig. 6c). In the region of the rhopalium, some neurons turned radially to extend into a compressed network within the smooth muscle band. Other neurons turned in the opposite direction and entered the rhopalial stalk.
The FMRFamide-IR network of the smooth muscle band (Fig. 6c) was less dense than the parallel tubulin-IR network. FMRFamide-IR neurons did not follow muscular outcroppings or travel into the subumbrella (Fig. 6c, inset). As the neuron tract neared the bell apex, it deviated from the radial axis and appeared to join the FMRFamide-IR nerve net in the manubrium. Here, the distribution of neurons exhibited a diffuse netlike arrangement (Fig. 7b). The [alpha]-tubulin-IR and the FMRFamide-IR nerve nets exhibited roughly equal distributions in the manubrium, with the former appearing slightly more dense than the latter.
Pedalia and tentacles
The primary structures for prey capture and delivery to the manubrium are the pedalia and tentacles. From the rhopalial niche, the nerve ring traveled obliquely downward to the bases of the bladelike pedalia located at the bell interradii (Fig. 8b). [alpha]-Tubulin-IR neurons descended from the nerve ring into an immunoreactive nerve net in the pedalium (Fig. 8a). The nerve net was diffuse, much like that of the subumbrella and velarium, except for a group of neurons that appeared to be associated with a bundle of longitudinal muscle fibers on the oral margin of each pedalium (Fig. 8a; Satterlie et al, 2005). Neurons from the pedalia extended into the tentacles and appeared to be deep to the epithelium, in the region of the longitudinal muscle.
At the interradii, FMRFamide-IR neurons entered the pedalia at two points on either side of the nerve ringpedalial junction (Fig. 8b). Clusters of immunoreactive cell bodies were found in this region and gave rise to a network that ran into the single tentacle of Carybdea or divided into each of the three pedalia and tentacles of Tripedalia (Fig. 8b). The FMRFamide-IR network was more superficial than the tubulin-IR nerve net and included stained epithelial sensory cells (Fig. 9).
A single, [alpha]-tubulin-IR exumbrellar network was found in Tripedalia (Fig. 10). The nerve net was of low density compared to the subumbrellar motor networks. Individual neurites were 2.6 [+ or -] 0.05 [micro]m in diameter, with sparse puncta. Neurites ran between nematocyst batteries that stained brightly with the tubulin antibodies. Due to this staining, it was not clear if the neurites contacted nematocytes. No FMRFamide immunoreactivity was present in the exumbrella of Tripedalia or Carybdea.
The peripheral network of the swim system
Each muscle sheet in the swim system of cubomedusae serves a different role in effective locomotion. The subumbrella generates the majority of propulsive force for fluid ejection from the bell. In this region, neurite orientations are independent of muscle fiber orientation and seem to form a "classical" cnidarian nerve net.
The velarium is an annular flap of swim muscle extending to the bell opening that serves to shape the aperature during a swim contraction, much like the velum in hydromedusae (Gladfelter, 1972, 1973). The velarial nerve net is similar to the subumbrella in its spatial organization of neurites, but it exhibits a greater network density.
Neurites appear to occasionally travel in bundles, at least in larger cubomedusae such as adult Carybdea, and thus our measurements of neurite density in Carybdea may be underestimates, and measurements of neuron diameter, overestimates. Nonetheless, relative differences in neurite density between muscle sheets are still relevant for this species, as bundling appears universal in the swim system. In Tripedalia, neurites did not appear to travel in bundles given that individual puncta were clearly visible along individual neurites of all three subumbrellar regions.
Asymmetric contractions contour the velarium to facilitate directional fluid ejection. This response can be induced using a point light source in Tripedalia (Petie et al., 2011). The velarial quadrant adjacent to the light source contracts prematurely and more strongly than other portions of the velarium, which exhibit delayed, weaker contractions. The result is an asymmetrical bell aperture through which water is preferentially expelled in a direction opposite the strongly contracting velarial quadrant, driving the medusa into a turn.
The increase in neurite density may provide velarial motor effectors with the necessary innervation field to elicit early, strong contractions, but the asymmetry of contractions presents another problem. The velarial nerve net is continuous around the bell. Signals from one rhopalium could easily travel from one bell quadrant to the next through the nerve ring and through the nerve net. A second subumbrellar nerve net, much like the scyphomedusan Diffuse Nerve Net, which stains with the anti-FMRFamide antibody, is not present in cubomedusae. This removes a potential source of peripheral variability since the scyphozoan Diffuse Nerve Net is capable of modulating swim muscle contractility (Horridge, 1956a, b; Passano, 1973). This raises a question about how the asymmetrical contractions of the velarium are accomplished.
One possible solution involves the perradial frenula, which possess the greatest motor network density of the swim muscle sheets in both Tripedalia and Carbydea. The frenular network is also unique in that neurites have a preferential orientation that is nearly parallel to the radially arranged muscle fibers.
The frenulum appears to be seamlessly integrated into the overall excitation system of the circular muscle sheets. Velarial muscle fibers turn 90[degrees] into the frenulum (Satterlie et al, 2005), and neurites from the velarium and nerve ring compress into the frenular network. The frenulum was thought to serve a structural role (buttress) in stabilizing the velarium during a swim contraction (Gladfelter, 1973), yet its position at the perradius just opposite the rhopalial niche is well suited to receive direct motor commands from the adjacent rhopalium. The anatomical evidence presented here, in concert with results of previous studies (Satterlie et al., 2005; Petie et al., 2011) suggests that the frenulum may serve a significant role in turning behavior.
Hydromedusae possess an independently derived, analogous structure to the velarium. Called the velum, it serves the same function as the velarium, a possible case of functional convergence. However, exumbrellar radial muscles in the hydromedusan velum contract locally to create asymmetric contractions (Gladfelter, 1972, 1973). This results in a deformed bell aperture that directionally expels water from the bell cavity. With the exception of the frenula, radial muscle is not associated with the cubomedusan velarium.
The arrangement of a dense network of neurons that run predominantly parallel to muscle fibers has implications for conduction velocity and facilitation properties of the nerve net. In the retractor muscles of the sea anemone Calamactis praelongus, conduction velocity measurements are fastest (~0.9 m/s) in the upper portions of the retractor muscle where the network is most dense and neurons conform to the direction of muscle fibers (Pickens, 1969). In addition, the facilitation rate of the musculature is greater in this region. The organized network of the cubomedusan frenulum strongly resembles that of the Calamactis retractor network, which suggests that conduction velocity and facilitation properties may be similarly high.
Neurite morphology in the swim system
Distinct punctate labeling of neurons is characteristic of cnidarian nerve nets. For example, Parkefelt and Ekstrom (2009) observed punctate labeling in FMRFamide-IR neurons in Carbydea marsupialis. In all species utilized in this investigation, similar observations were made using antibodies targeting FMRFamide-IR and a-tubulin-IR neurons. Punctate immunolabeling patterns suggest the presence of multiple synaptic sites along the lengths of neurites, but the morphological and physiological role played by these puncta is not known.
Pantin (1935) first suggested that synapses occur along the lengths of neurons in cnidarians, and punctate labeling patterns support this argument. The very large puncta at intersections of neurons in the velarium suggest points of reciprocal synapse communication, as described in scyphomedusae by Anderson (1985). Punctate labeling is less visible in the frenula, but this is likely due to the sheer density of neurons that obscure them from easy viewing.
Advantages of a diffuse peripheral network
Nerve nets are considered a predominant feature in the nervous systems of cnidarians and other basal metazoans, leading to assumptions concerning the primitive design in these organisms (Satterlie, 2011). Hydromedusae rely on epithelial conduction, with or without nerve nets, for spreading excitation through their two-dimensional sheets of swim muscle (Spencer, 1978, 1979; Satterlie and Spencer, 1983; Mackie, 2004). Cubomedusae and scyphomedusae appear to lack gap junctions (Mackie et al., 1984; Satterlie, 2011) and thus must innervate broad sheets of muscle cells via chemical synapses. This raises questions about excitation efficiency in the context of radial symmetry and the distributed nature of swim pacemakers (in the rhopalia). A diffuse network of randomly oriented neurons, as two-dimensional as the muscle sheet it innervates, must be able to conduct excitation from multiple originating sites around the margin of the bell (non-polarized network), as seen with multiple marginal rhopalia, each with the ability to initiate a swim contraction. These necessary properties suggest that a diffuse nerve net organization may be the most efficient means of coordinating swim contractions in the absence of gap junctions and epithelial conduction.
Brightly stained FMRFamide-IR cells can be found in the pacemaker region of each rhopalium and at the base of each pedalium. Since rhopalia develop from polyp tentacles and future medusoid tentacles develop de novo between the developing rhopalia (Werner et al., 1971; Straehler-Pohl and Jarms, 2005), an interesting question is whether the FMRFamide-IR clusters of the rhopalia and pedalia have a similar developmental origin.
The FMRFamide-IR network of the rhopalium did not include the full complement of RFamide-IR networks described by Parkefelt and Ekstrom (2009). But it did include a pad of putative epithelial sensory cells in the superior, posterior wall of the rhopalium. Processes of this suspected sensory pad joined the stained network of neurons in the pacemaker region, a finding not reported by Parkefelt and Ekstrom (2009) or Skogh et al. (2006). RFamide-IR sensory cells were described in the lateral walls of the rhopalia (Parkefelt and Ekstrom, 2009; Skogh et al., 2006), but without description of connections with the rhopalial FMRFamide-IR neuronal network. This suggests that the rhopalia may have multiple sensory pads on its posterior and lateral walls, and that the FMRFamide-IR network may have a sensory role in the rhopalium.
Similar sensory pads, termed touch plates, have been described from the rhopalia of scyphomedusae (Spangenberg et al., 1996; Nakanishi et al., 2009). FMRFamide-IR sensory cells have been verified in the rhopalia and rhopalial niches of scyphomedusae (Satterlie and Eichinger, 2014). Since the statolith of both scyphozoan and cubozoan rhopalia is not organized into a recognizable statocyst-like structure, it is possible that the statolith functions as a weight to maintain a specific rhopalial orientation independent of orientation of the medusa, as demonstrated by Garm et al. (2011). The rhopalial touch plates and sensory pads may represent sensory components of statocyst function in cubomedusae considering the tight placement of rhopalia in their rhopalial niches (which can totally enclose the rhopalia in some species) and the constant movement of the tissues during swimming. The sensory cells and accompanying neural network of the rhopalial niches may be included in this function.
Differences in staining presumably result from use of different antibodies; a commercial FMRFamide antibody in our work and that of Skogh et al. (2006), and an RFamide antibody in other studies (Martin, 2004; Parkefelt and Ekstrom, 2009). Several RFamides have been localized in cnidarian tissues (Grimmelikhuijzen et al., 1996, 2002), so it is likely the FMRFamide antibody exhibited more restricted staining compared to the RFamide antibody. This is interesting since it suggests there may be multiple peptidergic networks in the rhopalia, and likely the subumbrella, of cubomedusae.
Peripheral networks--coordination of feeding and defense
Bundles of [alpha]-tubulin-IR and FMRFamide-IR neurons in the smooth muscle bands represent potential pathways for coordinating perradial smooth muscle contractions and manubrial movements. The smooth muscle bands serve several functions in cubomedusae such as contributing to directional orientation of the manubrium during feeding and in pulling the tentacles into the bell cavity during feeding and crumple responses. The latter is a defensive response employed by many jellyfish, including cubomedusae, in which harsh external stimuli cause a cessation of feeding and swimming, and an infolding of the bell margin and tentacles (Hyman, 1940). Scemes and McNamara (1991) suggest that radial muscle and longitudinal muscle are responsible for retraction of the bell margin and tentacles, respectively, in hydromedusae. In cubomedusae, the longitudinal muscle produces tentacle and pedalial retraction, whereas the subumbrellar radial muscle bands are active during feeding and crumpling behaviors.
The quadrangular manubrium hangs from the bell apex where the subumbrellar smooth muscle bands coalesce to form its perradial ridges. Neurons in the manubrial body form a nerve net similar in density and appearance to that found in the subumbrella. In Chironex fleckerii, the manubrium is capable of directly capturing and restraining small segestid shrimp (Acetes australis) that swim directly into the subumbrellar cavity without touching any tentacles. If the shrimp touch the walls of the subumbrellar cavity, the manubrium moves to that bell quadrant and the lips grasp the shrimp for ingestion (Hamner et al., 1995). In the normal feeding response, the manubrium orients toward an inward-bending pedalium with tentacle-ensnared prey (Larson, 1976). On occasion the manubrium moves to intercept prey prior to inward bending of the pedalium or before the prey items even enter the bell cavity (Hamner et al., 1995). Collectively this behavior suggests very specific asymmetric control of manubrial activity that can be independent of tentacular activity.
The pedalia and tentacles are densely innervated by [alpha]-tubulin-IR and FMRFamide-IR nerve nets (Satterlie, 2011). In the tentacles, the tubulin-IR nerve net is deep to the epithelium, suggesting a motor function for the longitudinal muscle of the tentacles. The FMRFamide-IR nerve net is located at the base of the epithelium, superficial to muscle cell bodies, and the inclusion of stained epithelial cells with surface projections suggests a sensory function for the tentacular network. Given their location, these sensory cells could potentially serve a varied assortment of sensory functions given the important role of the tentacles in feeding and defense (including nematocyst discharge) and in stabilizing the bell while swimming (Shorten et al., 2005).
A single tubulin-IR nerve net was found in the exumbrella of Tripedalia. Exumbrellar nerve nets are found in a number of scyphomedusae, where both tubulin-IR and FMRFamide-IR networks are present (Satterlie and Eichinger, 2014). Physiological connectivity exists between the subumbrellar and exumbrellar diffuse nerve nets in Cassiopea (Passano, 2004), suggesting that at least the exumbrellar FMRFamide-IR network may be part of the Diffuse Nerve Net in scyphomedusae.
The tubulin-IR exumbrellar network in Tripedalia may serve a stimulatory role, coordinating firing of nematocyst batteries. Exumbrellar neurons appear to run between batteries, suggesting a coordinating or modulating role in nematocyte activity. It is curious that an FMRF-amide-IR network is missing in the exumbrella, since peptidergic innervation of nematocysts is apparently a common organization pattern throughout the Cnidaria (Anderson et al., 2004; Anderson and Bouchard, 2009). This is a point of distinction between cubomedusae and scyphomedusae since the latter group exhibits both tubulin-IR and FMRFamide-IR exumbrellar networks (Satterlie and Eichinger, 2014). A second exumbrellar network in the cubomedusae may yet exist, but lack immunoreactivity to our anti-FMRFamide antibody.
At least one carybdeid, Copula sivickisi (formerly Carybdea sivickisi), would benefit from dual nerve nets, or a cooptive nerve net. This species has four adhesive pads on the exumbrella near the apex, with which it attaches to algal substrates during the day (Hartwick, 1991). It would be interesting to know if more than one exumbrellar network is required for this species to accomplish "docking" behavior.
Current data suggest that the tubulin-IR motor nerve net is the sole source of innervation for the swim musculature (Fig. 1 la). Neurons immunoreactive to FMRFamide are not present throughout the subumbrella (Fig. 1 lb), so a Diffuse Nerve Net, as seen in scyphomedusae, may not exist in cubomedusae. This does not rule out the possibility that other RFamides, or other peptides, may provide widespread modulatory input to the swim musculature.
In the cubomedusan swim system, the FMRFamide-IR system is no longer peripheral in the swim musculature as in scyphomedusae. Instead, it is limited to more compressed networks in the rhopalia and at the pedalial-nerve ring junctions, connected by stained tracts in the nerve ring (Fig. 1 lb). In this way, the peripheral nature of the scyphomedusan Diffuse Nerve Net appears to be reduced and centralized in the cubomedusan swim system.
The presence of FMRFamide-IR immunoreactive neurons in the peripheral ganglia and visual systems of cubomedusae (Satterlie, 2002; Martin, 2002, 2004; Plickert and Schneider, 2004; Skogh et al., 2006, Parkefelt and Ekstrom, 2009) suggests an important role in photoreception. Peripheral structures such as the manubrium, pedalia, and tentacles also possess FMRFamide-IR nerve nets that include putative sensory cells, representing a consistency in the results from both scyphomedusae and cubomedusae. This suggests that the tubulin staining networks (compressed and diffuse) may have a primarily motor function, whereas the FMRFamide-IR networks (compressed and diffuse) have a sensory function coupled with a possible secondary motor function. The most dramatic difference is the possible lack of a peripheral Diffuse Nerve Net system in cubomedusae, which instead have a more centralized system represented by the FMRFamide-IR rhopalial network with its sensory pad.
We thank Dr. Alina Szmant for collecting specimens, and Drs. R. Dillaman and S. Pyott for comments on the written work. This project was supported by NSF grant IOS-0920825 and the Frank Flawkins Kenan Endowment (to RAS), and by a grant from Sigma Xi to JE. The monoclonal tubulin antibody was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICFID and maintained by The University of Iowa, Department of Biology, Iowa City, IA 52242.
Received 18 March 2013; accepted 4 February 2014.
Anderson, P. A. V. 1985. Physiology of a bi-directional, excitatory, chemical synapse. J. Neurophysiol. 53: 821-835.
Anderson, P. A. V., and C. Bouchard. 2009. The regulation of cnidocyte discharge. Toxicon 54: 1046-1053.
Anderson, P. A. V., and W. E. Schwab. 1981. The organization and structure of nerve and muscle in the jellyfish Cyanea capillata (Coelenterata; Scyphozoa). J. Morphol. 170: 383-399.
Anderson, P. A. V., A. Moosler, and C. J. Grimmelikhuijzen. 1992. The presence and distribution of antho-RFamide-like material in scyphomedusae. Cell Tissue Res. 267: 67-74.
Anderson, P. A. V., L. F. Thompson, and C. G. Moneypenny. 2004. Evidence for a common pattern of peptidergic innervation of cnidocytes. Biol. Bull. 207: 141-146.
Coates, M. M. 2003. Visual ecology and functional morphology of the Cubozoa. Integr. Comp. Biol. 43: 542-548.
Garm, A., M. M. Coates, R. Gad, J. Seymour, and D.-E. Nilsson. 2007a. The lens eyes of the box jellyfish Tripedalia cystophora and Chiropsalmus sp. are slow and color blind. J. Comp. Physiol. A 193: 547-557.
Garm, A., Y. Poussart, L. Parkefelt, P. Ekstrom, and D.-E. Nilsson. 2007b. The ring nerve of the box jellyfish Tripedalia cystophora. Cell Tissue Res. 329: 147-157.
Garm, A., M. Oskarsson, and D.-E. Nilsson, 2011. Box jellyfish use terrestrial visual cues for navigation. Curr. Biol. 21: 798-803.
Gladfelter, W. B. 1972. Structure and function of the locomotory system of Polyorchis montereyensis (Cnidaria, Hydrozoa). Helgol. Mar. Res. 23: 38-79.
Gladfelter, W. G. 1973. A comparative analysis of the locomotory systems of medusoid cnidaria. Helgol. Wiss. Meeresunters 25: 228-272.
Grimmelikhuijzen, C. J. P. 1985. Antisera to the sequence Arg-Pheamide visualize neuronal centralization in hydroid polyps. Cell Tissue Res. 241: 171-182.
Grimmelikhuijzen, C. J. P., I. Leviev, and K. Carstensen. 1996. Peptides in the nervous systems of cnidarians: structure, function and biosynthesis. Int. Rev. Cytol. 167: 37-89.
Grimmelikhuijzen, C. J. P., M. Williamson, and G. N. Hansen. 2002. Neuropeptides in cnidarians. Can. J. Zool. 80: 1690-1702.
Hamner, W. M., M. S. Jones, and P. P. Hamner. 1995. Swimming, feeding, circulation and vision in the Australian box jellyfish, Chironex fleckeri (Cnidaria, Cubozoa). Mar. Freshw. Res. 46: 985-990.
Hartwick, R. F. 1991. Observations on the anatomy, behaviour, reproduction and life cycle of the cubozoan Carybdea sivickisi. Hydrobiologia 216/217: 171-179.
Horridge, G. A. 1956a. The nerves and muscles of medusae. V. Double innervation. J. Exp. Biol. 33: 366-383.
Horridge, G. A. 1956b. The nervous system of the ephyra larva of Aurellia aurita. Q. J. Microsc. Sci. 97: 59-74.
Hyman, L. H. 1940. The Invertebrates, Vol. 1, Protozoa Through Ctenophora. McGraw-Hill, New York. 726 pp.
Larson, R. J. 1976. Cubomedusae: feeding, functional morphology, behaviour, and phylogenetic position. Pp. 237-245 in Coelenterate Ecology and Behaviour, G. O. Mackie, ed. Plenum Press, New York.
Mackie. G. O. 2004. Central neural circuitry in the jellyfish Aglantha. Neurosignals 13: 5-19.
Mackie, G. O., P. A. V. Anderson, and C. L. Singla. 1984. Apparent absence of gap junctions in two classes of Cnidaria. Biol. Bull. 167: 120-123.
Martin, V. J. 2002. Photoreceptors of cnidarians. Can. J. Zool. 80: 1703-1722.
Martin, V. J. 2004. Photoreceptors of cubozoan jellyfish. Hydrobiologia 530/531: 135-144.
Matsumoto, G. I. 1995. Observations on the anatomy and behavior of the cubozoan Carybdea rastoni Haacke. Mar. Freshw. Behav. Physiol. 26: 139-148.
Nakanishi, N., V. Hartenstein, and D. K. Jacobs. 2009. Development of the rhopalial nervous system in Aurelia sp. 1 (Cnidaria, Scyphozoa). Dev. Genes Evol. 219: 301-317.
Pantin, C. F. A. 1935. The nerve net of the Actinozoa. II. Plan of the nerve net. J. Exp. Biol. 12: 139.
Parkefelt, L., and P. Ekstrom. 2009. Prominent system of RFamide immunoreactive neurons in the rhopalia of box jellyfish (Cnidaria: Cubozoa). J. Comp. Neurol. 516: 157-165.
Passano, L. M. 1965. Pacemakers and activity patterns in medusae: homage to Romanes. Am. Zool. 5 : 465-481.
Passano, L. M. 1973. Behavioral control systems in medusae: a comparison between hydro- and scyphomedusae. Publ. Seto Mar. Biol. Lab. 20: 615-645.
Passano, L. M. 2004. Spasm behavior and the diffuse nerve-net in Cassiopea xamachana (Scyphozoa: Coelenterata). Hydrobiologia 530/ 531: 91-96.
Petie, R., A. Garm, and D.-E. Nilsson. 2011. Visual control of steering in the box jellyfish Tripedalia cystophora. J. Exp. Biol. 214: 2809-2815.
Pickens, P. E. 1969. Rapid contractions and associated potentials in a sand-dwelling anemone. J. Exp. Biol. 31: 513-528.
Plickert, G., and B. Schneider. 2004. Neuropeptides and photic behavior in Cnidaria. Hydrobiologia 530/531: 49-57.
Reynolds, E. S. 1963. The use of lead citrate at high pH as an electronopaque stain in electron microscopy. J. Cell Biol. 17: 208-212.
Romanes, G. J. 1885. Jellyfish, Starfish and Sea Urchins, Being a Research on Primitive Nervous Systems. D. Appleton, New York.
Satterlie, R. A. 2002. Neural control of swimming in jellyfish: a comparative story. Can. J. Zool. 80: 1654-1669.
Satterlie, R. A. 2011. Commentary: Do jellyfish have central nervous systems. J. Exp. Biol. 214: 1215-1223.
Satterlie, R. A., and J. M. Eichinger. 2014. Organization of the ectodermal nervous structures in jellyfish: Scyphomedusae. Biol. Bull. 226: 29-40.
Satterlie, R. A., and A. N. Spencer. 1983. Neural control of locomotion in hydrozoan medusae. A comparative study. J. Comp. Physiol. A 150: 195-205
Satterlie, R. A., K. S. Thomas, and G. C. Gray. 2005. Muscle organization of the cubozoan jellyfish Tripedalia cystophora Conant 1897. Biol. Bull. 209: 154-163.
Scemes, E., and J. C. McNamara. 1991. The ultrastructure of the radial neuromuscular system of the jellyfish Liriope tetraphylla (Hydrozoa, Trachymedusae): implications in crumpling behavior. Biol. Bull. 181: 474-483.
Schwab, W. E., and P. A. V. Anderson. 1980. Intracellular-recordings of spontaneous and evoked electrical events in the motor neurons of the jellyfish Cyanea capillata. Am, Zool. 20: 941.
Shorten, M. O., J. Davenport, J. Seymour, M. C. Cross, T. J. Carrette, G. Woodward, and T. F. Cross. 2005. Kinematic analysis of swimming in Australian box jellyfish, Chiropsalmus sp. and Chironexflecked (Cubozoa, Cnidaria, Chirodropidae). J. Zool. (Lond) 267: 371-380.
Skogh, C., A. Garm, D.-E. Nilsson, and P. Ekstrom. 2006. Bilaterally symmetrical rhopalial nervous system of the box jellyfish Tripedalia cystophora. J. Morphot. 267: 1391-1405.
Spangenberg, D., E. Coccaro, R. Schwarte, and B. Lowe. 1996. Touch-plate and statolith formation in graviceptors of ephyrae which developed while weightless in space. Scanning Microsc. 10: 875-888.
Spencer, A. N. 1978. Neurobiology of Polyorchis: I. Function of effector systems. J. Neurobiol. 9: 143-157.
Spencer, A. N. 1979. Neurobiology of Polyorchis: II. Structure of effector systems. J. Neurobiol. 10: 95-117.
Spurr, A. R. 1969. A low-viscosity epoxy resin embedding medium for electron microscopy. J. Ultrastruct. Res. 26: 31-43.
Stewart, S. 1996. Field behavior of Tripedalia cystophora (Class Cubozoa). Mar. Freshw. Behav. Physiol. 27: 175-188.
Straehler-Pohl, L., and G. Jarms. 2005. Life cycle of Carybdea marsupialis Linnaeus, 1758 (Cubozoa, Carybdeidae) reveals metamorphosis to be a modified strobilation. Mar. Biol. 147: 1271-1277.
Werner, B., C. E Cuttress, and J. P. Studebaker. 1971. Life cycle of Tripedalia cystophora Conant (Cubomedusae). Nature 232: 582583.
Werner, B., D. M. Chapman, and C. E. Cutress. 1976. Muscular and nervous systems of the cubopolyp (Cnidaria). Experientia 32: 10471049.
JUSTIN M. EICHINGER, AND RICHARD A. SATTERFIE *
Department of Biology and Marine Biology and Center for Marine Science, University of North Carolina Wilmington, Wilmington, North Carolina 28409
* To whom correspondence should be addressed. E-mail: satterlier@ uncw.edu
|Printer friendly Cite/link Email Feedback|
|Author:||Eichinger, Justin M.; Satterlie, Richard A.|
|Publication:||The Biological Bulletin|
|Date:||Feb 1, 2014|
|Previous Article:||Organization of the ectodermal nervous structures in jellyfish: scyphomedusae.|
|Next Article:||Identifying the cellular mechanisms of symbiont-induced epithelial morphogenesis in the squid-vibrio association.|