Burrowing behavior in mud and sand of morphologically divergent polychaete species (Annelida: Orbiniidae).
Marine sediments account for about 70% of the Earth's biome and provide a habitat for diverse fauna. Burrowing organisms, such as worms, play important roles in the ecosystem, for example, through bioturbation the mixing of sediments and their pore waters which impacts nutrient cycling and alters sediment structure. These burrowers have been referred to as ecosystem engineers because they modify the physical structure of their habitat (Meysman et al., 2006). Muddy and sandy sediments differ both in their micro and macrofaunal inhabitants (e.g., Sanders, 1958: Herman et al., 2001) and in physical properties (e.g., Boudreau et al., 2005; Dorgan et al., 2006).
Marine muds, in which cohesive organic material holds mineral grains together, behave elastically, and worms burrow by cracking the material. Burrowers exert dorsoventral forces against the elastic burrow walls, concentrating stressat the anterior burrow tip, which extends anteriorly by fracture (Dorgan et al., 2005, 2007). Worms have been shown to apply these dorsoventral forces either by using an eversible pharynx or proboscis (Dorganet al., 2007; Murphy and Dorgan, 2011) or by anterior expansions using their hydrostatic skeletons (Che and Dorgan, 2010).
Sand is a non-cohesive granular material and thus differs mechanically from mud (Boudreau et al., 2005): stress in muds is transmitted through the elastic organic matrix; in sands, grains rest on each other and the weight of overlying sediment is transmitted along particle contacts, with some bearingdisproportionately higher or lower amounts of the load (Gong et al., 2001). Non-uniform distributions of stress across numerous particles form patterns referred to as stress chains. Since the mechanics of sands depend on the gravitational forces acting on individual grains rather than adhesion of the organic material, the physical behavior depend on particle size, surface characteristics, heterogeneity of particles, and packing structure (Duran, 2000; Goldenberg and Goldhirsch, 2005). Burrowers in sands experience different stress distributions than those in muds, and they are not able to extend burrows by crack propagation because sand is not cohesive. Lack of cohesion and non-uniform stress distribution may also lead to burrow collapse once the worm moves through the burrow and no longer applies direct force to burrow walls. Backward movements observed in worms burrowing in gelatin, an analog for muds (e.g., Che and Dorgan, 2010), may result in burrow collapse in natural sediments, although cohesion makes burrow collapse less likely in muds than in sands. Some worms, however, excrete mucus that can glue particles together and reduce the likelihood of burrow collapse. Because of these mechanical differences between muds and sands, we expect burrowing animals to utilize different burrowing strategies in the two materials, potentially includingboth behavioral and morphological differences.
Burrowers comprise taxa from many different phyla: even within annelids alone, considerable taxonomic and morphological diversity exists. There are over14,000 described species of polychaetes grouped into about 80 families (Rouse, 2007), including a diversity of feeding guilds and locomotory strategies Fauchald and Jumars, 1979). Differences among species may result from adaptations to different environments or may simply reflect different evolutionaryhistories. Functionally important morphologies often show evolutionary convergences, yet the phylogeny for annelids remains in flux (see Rousset et al.,2007; Zrzavy et al., 2009; Struck et al., 2011), making identifying these convergences challenging.
To simplify this problem, we chose to compare morphologies and burrowing behaviors of species within one family in which the phylogenetic relationships are well understood and that live in different habitats. The advantage to this approach is that we can identify morphological and behavioral differences that are evolutionarily divergent. Even within polychaete families, there can be considerable morphological (e.g., Opheliidae; Law et al.. 2014) and behavioral (e.g., Nereididae; Fauchald and Jumars, 1979) diversity or, in some cases (e.g., Nephtyidae), remarkably little morphological diversity.
To address the question of how habitat (mud vs. sand) affects behavior and is reflected in distinct morphological features of burrowers, we focused on three species from the family Orbiniidae-Leitoscoloplos pugettensis (Pettibone 1957), Naineris dendritica (Kinberg 1867), and Orbinia johnsoni (Moore 1909)-that inhabit either muds (L. pugettensis) or beach sands (the latter two species) and exhibit several morphological differences (Fig. 1; Table 1). Species in Orbiniidae are found worldwide in shallow to deep marine sediments. All orbiniids have been described as motile deposit feeders that do not make permanent burrows (Fauchald and Jumars, 1979). They lack anterior appendages and have an anterior muscular thoracic region followed by a more fragile abdomen (Rouse and Pleijel, 2001). Morphologically these three species differ primarily in anterior head shape and parapodial size (Table 1). Both L. pugen-tensis (Fig. 1A) and O. Johnsoni (Fig. 1B) have pointed anteriors with shorter parapodia compared to body width. O. Johnsoni has additional subpodial lobes, or papillae, in the posterior thoracic and anterior abdominal regions (Rouse and Pleijel, 2001). N. dendritica (Fig. 1C) has a broad, shovel-shaped anterior and longer external parapodia compared to body width (less visible in anesthetized worms (Fig. 1C) but apparent during burrowing). Molecular phy-logeny shows that these three species represent the three largest Glades within Orbiniidae and that L. pugettensis and O. Johnsoni are more closely related to each other than to N. dendritica (Bleidorn et al., 2009).
Morphological and behavioral differences among the three species of orbiniidae Naineris dendritica Leitoscoloplos pugettensis Orbinin johnsoni Broad, shovel-shaped Narrow pointy head Narrow pointy head head Lives in sand Lives in mud Lives in sand Long parapodia to Short parapodia to Short parapodia to body width body width body width No ventral papillae No ventral papillae Ventral papillae No active twisting Active twisting Active twisting No internal body No internal body Internal body movement movement movement
We examined burrowing behaviors of these three species in both mud and sand analogs, then related differences in morphologies and burrowing strategies to the differing material properties of their natural environments. While the morphological differences we identified are small relative to the tremendous morphological diversity of annelids, we hypothesize that they are likely to be functionally important in burrowing. We focused on the physical differences between muds and sands to explain the morphological, behavioral, and environmental differences among the three species
Materials and Methods
Leitoscoloplos pugettensis (width of 6th segment 0.98 [+ or -]0.15 mm; mean [+ or -] s.d.) was collected from mud flats at Fiesta Island in Mission Bay,San Diego, California, at low tide. Naineris dendritica (width of 6th segment 1.58 [+ or -] 0.21 mm; mean [+ or -] s.d.) was collected from the flowing seawater intake trough on Scripps Pier, Scripps Institution of Oceanography, La Jolla, California. N. dendritica has been found in the sandy beach sediments at Scripps Beach, but was much more abundant in the intake trough, which was lined with local sandy sediments. Orbinia johnsoni (width of 6th segment 1.06 [+ or -] 0.09 mm; mean s.d.) was collected from sandy sediments at Scripps Beach, La Jolla, California, at low tide. All animals were sorted and stored in individual containers with sediment from the collection site under flowing seawater until use in experiments. After use, animals were anesthetized in 7.5% Mg[Cl.sub.2] and preserved in 10% formalin.
Experimental setup for mud analog
Gelatin in seawater (28.35 g [liter.sup.-1]) was used as an analog for muddy sediments due to its similar mechanical properties (Johnson et al., 2002; Boudreau et al., 2005). Kinematics of burrowing by several species of polychaetes have been studied using gelatin as a transparent analog for muds (Dorgan et al., 2007; Che and Dorgan, 2010; Murphy and Dorgan, 2011).
The experimental setup for burrowing in gelatin was similar to that described previously (Che and Dorgan, 2010). Video was taken with a CCD camera (7.5 fps) (Basler A622f) with a left-handed circular polarizing filter. The aquarium was backlit using a photographic light table covered by a right-handed circular polarizing filter. Videos were recorded using LabView ver. 7.1.1 (National Instruments, Austin, TX). All gelatin experiments were conducted in an 11 [degrees]C cold room in order to maintain consistent material properties. Food-grade gelatin (Natural Foods Inc., Toledo, OH) was dissolved in seawater and boiled, poured into a glass aquarium 6.5 cm X 6.5 cm with a height of 9.5 cm, and allowed to cool overnight at 4 [degrees]C. Gelatin was always used the day after it was prepared to minimize bacterial growth and degradation of the material.
For burrowing experiments, a vertical crack was initiated in the gelatin in the middle of the aquarium and a worm was placed in the crack to begin burrowing. The top of the gelatin was covered in a thin layer of seawater to provide oxygen for the worm and maintain a gelatin-seawater interface. To encourage burrowing, the worm was occasionally prodded gently with a plastic pipette and directed into the crack. The lens on the video camera had a fixed focal length so the camera or aquarium was moved to keep the worm in focus while it burrowed.
Experimental setup for sand analogs
Two different materials were used to visualize burrowing in sandy sediments, cryolite and 500-[micro]m glass beads. The natural mineral cryolite ([Na.sub.3]Al[F.sub.6]) has been shown to have a refractive index of 1.338-1.339, similar to that of seawater, 1.339 (25 [degrees]C) (Josephson and Flessa, 1972). This makes cryolite a good analog to visualize burrowing in sand because it becomes nearly transparent when immersed in seawater. However, the cryolite must be from a natural worm was encouraged to burrow close to the acrylic plastic wall for the best visualization of burrowing behavior. We did not observe obvious differences between worms burrowing close to and farther from the wall but cannot eliminate the possibility of wall effects. Proximity was required, however, for proper visualization of burrowing behaviors. No polarizing filter was used, and experiments were conducted at room temperature with ambient light present. Cryolite was stored in a vial of seawater before and after experiments. mineral source because aluminum oxide impurities in synthetic cryolite decrease transparency and are toxic to animals (Josephson and Flessa, 1972). Cryolite was obtained as gravel-sized rocks (Ward Scientific) and was ground with a mortar and pestle, passed through a 500-[micro]m sieve, and then smoothed by rotation in a rock tumbler (Lortone, Inc., model 3A) overnight.
The cryolite video experiments followed a similar setup to the gelatin experiments. The aquarium used was narrower, 1.5 cm wide X 6.8 cm long x 8.5 cm high, and the worm was encouraged to burrow close to the acrylic plastic wall for the best visualization of burrowing behavior. We did not observe obvious differences between worms burrowing close to and farther from the wall but cannot eliminate the possibility of wall effects. Proximity was required, however, for proper visualization of burrowing behaviors. No polarizing filter was used, and experiments were conducted at room temperature with ambient light present. Cryolite was stored in a vial of seawater before and after experiments.
Glass beads have been used in experiments to visualize burrowing in a granular material (Winter et al., 2012). However, spherical particles such as glass beads have significantly different properties than more irregularly shaped particles such as sand grains, which have higher internal friction angles and lower tendency to rotate (Maeda et al., 2010). Parameters like peak friction angle are often lower in glass beads than in real soils due to the perfectly spherical shape of particles, which roll easily (Thomas and Bray, 1999). The uniformity of bead size results in lower packing density than in heterogeneous sands. These different properties could cause different mechanical responses of the media, for example, fluidization or burrow collapse, and possibly even different burrowing behaviors. Although cryolite was predominantly used in these experiments, experiments were also conducted in glass beads (using identical setups) to compare both material responses and the worm behaviors. Glass beads used were slightly larger than cryolite grains (500-[micro]m beads vs. <500-[micro]m cryolite grains) and were chosen primarily for visibility reasons; smaller beads substantially reduce visibility of burrowing.
Analysis of kinematics
We first identified general qualitative differences in burrowing behaviors among species and materials and described burrowing cycles for each species in each material. We measured distance traveled over time for the anterior end of the worm, from which we could evaluate periodicity of the burrowing cycle. Preliminary observations showed changes in anterior shape over a burrowing cycle for N. dendritica, which has a rounded anterior, so head width was measured for N. dendritica and for L. pugettensis, which has a pointed anterior. We also observed twisting behavior to determine if it occurred at regular intervals within burrowing cycles. We hypothesized that worms would exhibit differences in backward slipping during each cycle in muds versusin sands, and we calculated the backward slip from the distance traveled by the anterior. We measured cycle distances and frequencies to determine whether worms increase burrowing speed by increasing cycle distance or frequency.
First, dorsal view videos were used to measure distance traveled and changes in head width over time. Segments selected had at least two continuous burrowing cycles during which the worm exhibited minimal pausing, probing, or backing up. Perpendicular burrowing with respect to the camera was preferred, but if the worm angled less than 30[degrees] toward or away from the camera, the video was still used. Videos were rejected if they were out of focus or if the worm got too close to the aquarium walls (only in gelatin; about 1.5 cm from wall). Videos were then subsampled using LabView and the segments were analyzed using Im-ageJ ver. 1.44 with the plugin MtrackJ ver. 1.5 (Meijering et al., 2012) to track the position of the worms' heads at every frame.Image was used to measure the width of the head at each frame. Distance traveled was calculated from head position using custom Matlab ver. R2010b (MathWorks, Natick, MA) scripts, and was plotted with head width to assess how head width and position in the burrowing cycle were related. Frames in whichthe worms were twisting about the longitudinal body axis were identified and compared to the distance traveled to assess whether twisting behavior was periodic-that is, occurred at the same position within each burrowing cycle.
For measurements of backward slip, velocity, cycle distance, and frequency, videos were selected in which the worm was burrowing perpendicular to the camera but at any orientation(e.g., dorsal or lateral). Burrowing cycles were easily distinguished as a period of forward movement followed by a shorterbackward slipping. Cycle distance and frequency were calculated from the maximum distance traveled in each cycle. Velocity was calculated as the slope of a regression line through distance traveled as a function of time (through all data). To quantify backward slipping, we calculated a slip ratio (backward distance per burrowing cycle/forward distance per burrowing cycle) from maximum and minimum distances traveled for each burrowing cycle. Slip ratios for each burrowing cycle were then averaged over a burrowing segment.
To determine whether worms burrowed faster by increasing their cycle distance or by increasing cycle frequency, velocities of each individual were plotted as function of cycle distance and cycle frequency to obtain a regression coefficient for the two relationships. Data were log-transformed and a linear regression was tested against a null hypothesis of a zero slope. If, for example, worms increased their velocity by increasing cycle frequency while maintaining constant cycle distance, the slope (b) of velocity (U) plotted as a function of frequency (f) would be 1,f = a[U.sup.b] , and that of velocity as a function of distance, d = a[U.sup.b], would be 0 (see Quillin, 1999). Average velocity was calculated by taking a best-fit line through the graph of distance traveled over time. To calculate cycle distance, the average of each forward distance per burrowing cycle was taken. Frequency was calculated as the inverse of period, the average of the times between subsequent minimum and maximum positions in each burrowing segment, multiplied by 2 to obtain a full period.
While reviewing primary video, we observed that some worms exhibited apparently regular burrowing cycles with a fairly constant cycle period, whereas other worms appeared to burrow more erratically and cycles were harder to discern. To quantify the cycle period and its regularity, we used two methods.First, the Fourier transform of distance traveled, de-trended to show just fluctuations about the mean velocity, was calculated using a custom Matlab script. Amplitude was calculated from the discrete Fourier-transformed data and plotted as a function of frequency, then x and y positions of the peak were identified as the cycle frequency and amplitude, respectively. In cases in which there were multiple peaks of similar size, the peak with a frequency corresponding to the measured cycle period was selected. Cycle frequency and amplitude were then averaged across individuals for each species in each material. Second, the autocorrelation of the de-trended distance traveled over time was calculated to determine the strength of the periodicity of burrowing, which was averaged for each species in each material. Autocorrelation coefficient was calculated as the ratio of peak in unbiased autocorrelation at a lag of one cycle to the peak at a zero lag.
Descriptions of burrowing behaviors
In a normal burrowing cycle for all materials, each worm stretches its body forward to the previous maximum, extends the burrow anteriorly to reach a new maximum, expands the anterior, and then moves backward some distance. This sequence of events is similar to that of the cirratulid polychaete Cirriformia moorei, which has been described in four stages: forward stretching, anterior crack extension, anterior body thickening, and peristaltic wave progression (Che and Dorgan, 2010). Burrowing cycles can be characterized by cycle frequency and cycle distance.
Leitoscoloplos pugettensis in mud analog (gelatin). Four individuals of L. pugettensis burrowed in gelatin and fit the criteria for analysis of videos.In all materials tested, L. pugettensis exhibited lateral twisting behavior about the body axis while burrowing (Fig. 2; movie, http://www. biolbull.org/content/supplemental). In the analog for mud, IES natural material, L. pugeuensts stretcnes rorwara in a dorsoventrally compressed position (Fig. 2A), extending the crack anteriorly and reaching a maximum distance traveled (Fig. 2B), then moves backward while twisting its body (Fig. 2C-D). The head width does not change; variation in width occurs posterior to the head as the retrograde peristaltic wave progresses along the worm (Fig. 2A-D). Thetwisting behavior is periodic and occurs as the worm extends the crack anteriorly then moves backward (Fig. 2C, D; Fig. 3A).
Leitoscoloplos pugettensis in sand analog (cryolite and glass beads). Nine individuals of L. pugettensis burrowed in cryolite and fit the criteria for analysis of videos. In cryolite, the sand analog, while stretching forward. L. pugettensis sometimes twists periodically from the right to the left as the burrow anteriorly extends (Fig. 2E, F). As the peristaltic wave travels posteriorly, the anterior body thickness increases (Fig. 2G) then decreases (Fig. 2H; movie, http:// www.biolbull.org/content/supplemental). The worms exhibited lateral head movements in which they appeared to be probing the material (Fig. 2E-H). Twisting behavior occurred hut was inconsistent across burrowing cycles, with worms at least somewhat twisted most of the time(Fig. 3B). When comparing cryolite to glass beads (n = 4), inconsistent twisting behavior occurred most of the time, as in cryolite, and movements were similarly probing and jerky (Fig. 2I-L, Fig. 3C). The material showed greater differences, however, because the beads appeared to be semi-fluidized while L. pugettensis burrowed, and grain movement was more apparent than in cryolite (movie, http:// www.biolbull.org/content/supplemental).
Naineris dendritica in sand analog (cryolite and glass beads). Three individuals of N. dendritica burrowed in cryolite and lit the criteria for analysis of videos. In this analog for sand, the natural material for N. dendritica, the head is narrow as it moves forward (Fig. 4E; movie. http:// www.biolbull.org/content/supplemental), but as it reaches the peak of its forward movement and burrow extension (Fig. 4F), the head widens and moves backwardFig. 4G, H; Fig. 3E). N. dendritica does not actively twist during each burrowing cycle (Fig. 4E-H: Fig. 3E), although some worms exhibited a more gradual twisting over a period of many burrowing cycles. Like L. pugettensis, N. dendritica had a pattern of forward movement that was similar in glass beads (n = 2) and in cryolite (Fig. 3F; Fig. 41-L; movie, http://www.biolbull.org/content/supplemental). Worms moved at similar speeds in both materials and expanded the head width while moving backward (Fig. 3E, F). Around burrowing N. dendritica, as for L. pugettensis, the beads showed greater movementand appeared more fluidized than the cryolite (movie, http://www.biolbull.org/content/ supplemental).
Naineris dendritica in mud analog (gelatin). Six individuals of N. dendritica burrowed in gelatin and fit the criteria for analysis of videos (Fig. 4A-D). The burrowing behavior is similar to that in cryolite: the head is narrow as the worm moves forward, in this case extending the crack (Fig. 4C; movie, http://www.biolbull.org/content/supplemental), and then expands as it reaches the peak of its forward movement and moves backward (Fig. 3D; Fig.4A-D). The burrowing cycle is slower in gelatin (at a lower temperature) than in cryolite.
Orbinia johnsoni in sand analog (cryolite and glass beads). Five individuals of O. Johnsoni burrowed in cryolite and fit the criteria for analysis of videos. In this analog for sand, the natural material for O. Johnsoni, movement is very rapid, and it is hard to discern a clear burrowing cycle. Non-periodic twisting occurs during each cycle (Fig. 5E-H), but the twisting does not follow a discernable pattern across multiple cycles (Fig. 5H; movie,http://www.biolbull. org/content/supplemental). After a period of rapid burrowing, O. Johnsoni typically slows or stops and expands its entire anteriorregion (most of the thorax). During this expansion, a dark region visible through the body wall, possibly a blood sinus, moves anteriorly (Fig. 5I, J).This anterior body expansion was observed for all 6 of the 6 individuals observed and occurred every 3.56 [+ or -] 1.82 s (mean [+ or -] s.d.) of active burrowing segments analyzed. Worms typically exhibited little forward or backward movement during anterior body expansions (Fig. 3H; Fig. 51-L). The head width remains constant while burrowing (similar to that of L. pugettensis) and, for this reason, was not measured (Fig. 3G-I). Burrowing behavior of O. Johnsoni in glass heads (several observations of only one individual) was not as similar to behavior in cryolite as it was for the other two species (Fig. 31; Fig. 5M-P; movie, http://www. biolbull.org/content/supplemental). The pattern of movement showed more discernable cycles, and O. Johnsoniexhibited fewer internal body movements in glass beads than in cryolite (an average of 1 per 7.25 s in a 29-s segment of active burrowing analyzed for one individual) (Fig. 3I). However, movement of beads was much more apparent than in cryolite, similar to the other Iwo species (movie, http:// www.biolbul.org/content/supplemental).
Orbinia johnsoni in mud analog (gelatin). Only two out of six worms tested burrowed in gelatin. Four worms were unable or unwilling to burrow. Compared to its behavior in cryolite. O. Johnsoni exhibits a more regular visible peristaltic behavior in gelatin, where it extends forward (Fig. 5A), reaches a maximum forward distance (Fig. 5B), and slips backward (Fig. 5C-D; movie, http://www.biolbull.org/ content/supplemental). The internal body movement observed in cryolite (and once in glass beads) was not observed in gelatin. One individual worm exhibited twisting behavior that did not show any obvious periodicity, whereasrwekt re at 'no rshoarmarl fnr tt,a INtlicar tairtrrn (Pin 3G)Analysis of kinematics
The proportion of backward slipping varied significantly among species (two-way ANOVA; P = 0.009); however, we found neither a significant difference between slip values and materials (P = 0.68) nor an interaction between species and material (P = 0.08) (Fig. 6; Table 2). A multiple comparison teston a one-way ANOVA with data from the three materials combined showed that backward slip by N. dendritica was significantly lower than that of L. pugetten-sis and O. Johnsoni (P < 0.05), which were not significantly different from each other (P > 0.05). With combined data from the three materials, the average slip for L. pugettensis was 0.44 [+ or -] 0.19, N. dendritica 0.19 [+ or -] 0.16, and O. Johnsoni 0.52 (+ OR -) 0.21. Analysis of data for each material showed a significant difference between the slip values in cryolite of N. dendritica and the other two species, which did not differ from each other (ANOVA; P = 0.0025). No significant differences among species were found for gelatin or glass beads (ANOVA; P > 0.05), although sample sizes in glass beads were small. Analysis of data for each species showed significant differences among materials only for L. pugettensis in gelatin versus glass beads (ANOVA; P = 0.01).
Using data combined across species and materials, burrowing velocity is significantly correlated with the frequency of burrowing cycles ([r.sup.2] = 0.66, P = 0.039) (Fig. 7A; Table 2) with a regression equation of f = [1.380.sup.0.76] . The 95% confidence interval for the slope (exponent) includes 1 ([0.50 1.03]), indicating that up to 100% of the change in velocity could result from a frequency change. The relationship between velocity and cycle distance, d = 1.40[U.sup.0.19], is not statistically significant ([r.sup.2] = 0.024, P = 0.07; 95% CI [-0.09 0.47]) (Fig. 7B). Size of worms was not taken into account for this regression analysis. For most individual species and materials, either there was not enough data or the velocity ranges were small (Fig. 7), so statistical analyses were not done.
To compare movement patterns over burrowing cycles, we de-trended the distance traveled over time to obtain a plot of forward and backward movement over time, or movement fluctuations. Data were Fourier transformed to obtain cycle frequency and amplitude, which were averaged to show mean movement patterns of different species in different materials (Fig. 8A). The amplitude of this move-ment-how much worms moved back and forward in a cycle-showed significant differences among species (two-way ANOVA; P < 0.05) but not among materials (P = 0.19) (Fig. 8B). In cryolite and glass heads, L. pugettensis exhibited significantly higher amplitude than N. dendritica (ANOVA multiple comparison test; P < 0.05), but neither differed from O. Johnsoni, and species differences were not significant in gelatin. L. pugettensis and N. dendritica showed no significant differences in amplitude in the different materials, but the amplitude for O. johnsoni in cryo-lite differed significantly from the other two materials. The frequency of burrowing cycles was significantly different among both species and materials (P < 0.05). In cryoliteand gelatin, all three species exhibited significantly different frequencies (ANOVA multiple comparison; P < 0.05) (Fig. 8C), whereas differences in glass beads were not significant. For L. pugettensis and N. dendritica, frequency was significantly lower in gelatin than in the other two media (which lid not differ from each other).Experiments were conducted at a lower temperature in gelatin, however, which likely explains the lower frequencies. For O. Johnsoni, frequency was significantly higher in cryolite than in the other two media.
No significant differences in autocorrelation coefficient were found either among species or materials (two-way ANOVA; P > 0.05).This analysis was complicated, however, by the results for O. johnsoni in cryolite, for which the autocorrelation analysis showed a longer lag that appeared to correspond with the periodicity of pausing for anterior body expansion in addition to the short period of a burrowing cycle (Fig. 9). In graphs that showed this longer-period autocorrelation (including all 5 of the 5 video segments of O. Johnsoni burrowing in cryolite that were long enough to analyze), considerable variability in autocorrelation of subsequent peaks made the data unreliable, and they were not used.
Summary of burrowing kinematics measured, separated by species and material Measurement Material Leitoscoloplos Nuineris Orbinia pugettensis dendritica johnsoni Slip Cryolite 0.524 [+ or -] 0.285 [+ or 0.553 [+ 0.19 (n =10) -] 0.11 (n or -] =2) 0.35 (n =4) Gelatin 0.336 [+ or -] 0.133 [+ or 0.575 [+ 0.25 (n =4) - ]0.11 (n or -] =6) 0.21 (n =2) Beads 0.214 [+ or -] 0.279 [+ or 0.503 (n 0.12 (n =4) -] 0.33 (n = 1) =2) Velocity Cryolite 0.548 [+ or -] 0.174 [+ or 0.365[+ 0.32 (n =10) -] 0.05 (n or -] =2) 0.1 (n =4) Gelatin 0.218 [+ or 0.133[+ or 0.166 [+ -]0.14 (n =4) -] 0.03 (n or -] =6) 0.05 (n =2) Beads 1.04 [+ or -] 0.285[+ or 0.504 (11 0.35 (n =4) -] 0.21 (n = 1) =2) Cycle Cryolite 0.564 [+ or -] 0.375 [+ or 0.229 [+ distance 0.21 (n =10) -] 0.13 (n or -] =2) 0.07 n =4) Gelatin 0.783 [+ or 1.10 [+ or 0.322 [+ -]0.35 (n =4) -] 0.22 (n or -] =6) 0.02 (n =2) Beads 1.18 [+ or -] 0.366 [+ or 0.434 (n 0.22 (n =4) -] 0.23 (n = 1) =2) Cycle Cryolite 0.949 [+ or -] 0.506 [+ or 1.77 [+ frequency 0.27 (n =10) -] 0.10 (n or -] =2) 0.14 (n =4) Gelatin 0.272 [+ or -] 0.246 [+ or 0.474 [+ 0.06 (n =4) -] 0.24 (n or -] =6) 0.16 (n =2) Beads 0.890 .[+ or -] 0.681[+ or 1.18 (n = 0.22 (n =4) -] 0.20 (n 1) =2) All gelatin measurement were taken at a lower temperature than in cryolite and glass beads. Values are mean +s.d(n=number of individual worms).
Burrowing behaviors and relation to natural material
Leitoscoloplos pugettensis extends its burrow via crack propagation, consistent with previous descriptions of burrowing by mud-dwelling worms (e.g.,Cirriformia moorei: Che and Dorgan, 2010). As suggested for C moorei, its pointed head likely amplifies the stress applied by the wider body at the tip of the burrow, which then extends by fracture (Che and Dorgan, 2010). L. pugettensis differs from C. moorei, however, in exhibiting peiodic twisting when burrowing in gelatin. Twisting was also observed in cryolite and glass beads, but was less consistent and periodic (Fig. 3A-C). Twisting likely increases the force on the burrow walls, as worms are considerably wider than thick, and orienting the width dorsoventrally would increase the opening of the crack and displacement of the burrow walls and consequently the force applied to the material. This increased force could facilitateburrow extension by fracture in mud. Our preliminary observations showed that the morphologically similar Leitoscoloplos sp. from muddy sediments inLowes Cove, Walpole, Maine, exhibits periodic twisting following anterior crack extension similar to that of L. pugettensis when burrowing in gelatin (Dorgan, unpubl. data). Leitoscoloplos and Scoloplos species sampled by Bleidorn et al. (2009) form a clack that branches into eastern and western UnitedStates Glades; that similar behaviors were observed in both sub-clades suggests that this periodic twisting behavior is likely characteristic of thewhole Clade. L. pugettensis slips backward more than N. dendritica in a burrowing cycle (Fig. 6), but mud is an elastic material and burrow collapse is therefore less likely to be an issue than in non-cohesive sands. The work of fracture is about 10 times the elastic work to displace sediment to open theburrow (Dorgan et al., 2011), so re-opening the burrow after elastic recoil of mud is relatively inexpensive compared to creating a new burrow in sand.Cryolite and glass beads lack the elastic rebound of gelatin, and the more erratic behavior of L.pugettensis in those materials may be a response to the absence of the dorsoventral compression force it experiences in a crack-shaped burrow.
We hypothesized that the increased head width and larger parapodia would provide increased friction and reduce backward slipping for N. dendritica, and our data show a significantly lower slip ratio for this species compared to O. Johnsoni and L. pugettensis (Fig. 6). This difference appears to depend more on morphology than on the material in which the worm is burrowing, as we found no significant effect of material on backward slipping. As N. dendritica moves backward, its head expands in width, creating a larger surface area contacting the material, and therefore increasing frictional resistance to backward slipping. This is especially important in sand, the natural habitat of N. den-dritica, since substantial backward slipping could allowfor burrow collapse in this non-cohesive material. Decreasing slip may increase burrowing efficiency, as burrow collapse would increase the work to burrow.
Orbinia johnsoni burrows in sands like N. dendritica, but its morphology more closely resembles that of L. pugetten-sis. The ratio of backward slip to forward movement by O. Johnsoni was similar to that of L. pugettensis and greater than that of N. dendritica, consistent with the hypothesis that thewider head and larger parapodia of N. dead ritica are important in preventing slip. We had hypothesized that the ventral papillae on O. Johnsoni would help prevent slip back, but this was not supported by our data. However, the pattern of movement for O. Johnsoni in cryolite was very different from those of N. dendritica and L. pugettensis (Fig. 8A). Burrowing cycles with high amplitude, or greater forward and backward movements in a cycle, and lowefrequency, with more time between movements, as exhibited by L. pugettensis in gelatin (Fig. 8A), could potentially allow for burrow collapse in non-cohesive sands. The distance traveled in each burrowing cycle was much shorter for O. Johnsoni compared to other burrowers (Fig. 3). These high-frequency short movements may prevent burrow collapse in sand and also allow the worms to quickly probe ifferent paths in this heterogeneous material. Our study did at explicitly address the use of mucus to prevent burrow pllapse, but we observed no fragile tube-like structures of and grains held together with mucus around any worms nor bvious differences in mucus secretion among the three pecies. The anterior body expansion that is accompanied y movement of internal fluid appears to pack sand grains, which may also prevent burrow collapse and reduce resistance to forward movement. Similar periodic anterior body Kpansions were not observed in either N. dendritica or L. ugettensis, but they have been observed in the opheliid olychaete Thoracophelia mucronata, found in the same nvironment as O. Johnsoni (Dorgan, unpuhl. data). In ad-ition to expansions of the thorax, T. mucronata has a pecialized injector organ for expanding just the head re-ion, and we have suggested that these expansions compact to burrow walls when burrowing in sands (Law et al., 2014). These head expansions are similar to expansions ssociated with fracture in muds (Che and Dorgan, 2010) rid are focused in the anterior few segments rather than the ntire thorax (~10 setigers). During an anterior body ex-ansion, the head of O. Johnsoni does not move forward or ackward (Fig. 3H, 1). Interestingly, anterior body expan-ons were observed more frequently in cryolite than in lass beads, consistent with a greater need to apply force to configure grains in a material in which grains are less asily displaced. These expansions were not observed at all gelatins, where grain packing is not necessary and no miming would occur. It is also possible that O. Johnsoni annot exert adequate dorsoventral force against the elastic bound of gelatin to expand the thorax-worms in gelatin ppeared quite dorsoventrally compressed and most of them id not burrow at all.
Our comparison of burrowing behaviors in and the material responses of cryolite and glass beads showed several ifferences that indicate that results from burrowing exper-nents in glass beads should be interpreted carefully.
Whereas burrowing behaviors of two species, L. pugettensis nd N. dendritica, did not show obvious differences be-ween the two media, the third species, O. Johnsoni, exhib-ed several differences. The anterior expansion occurred tuch less frequently in glass beads than in cryolite, and the urrowing cycles were much easier to distinguish (Fig. 3G, 1). The material response differed for all three species: we oted much more grain movement around burrowers in the lass beads than among those in cryolite (movie, http:// www.biolbull.org/content/supplemental). Although we did of measure packing density to determine whether grains were fluidized, the differences in grain movement observed aggest that fluidization would occur much more readily in lass beads than in cryolite. Fluidization has been described s a burrowing mechanism in a variety of sand-dwelling ivertebrates (Emerita portoricensis, Ensis directus) and has been suggested to reduce the energetic cost of burrowing (Trueman, 1975). Observations of fluidization in glass beads (e.g., Winter et al., 2012) do not necessarily indicate fluidization of natural sandy sediments. Additionally, if fluidization reduces the energetic cost of burrowing (see Winter et al., 2012), worms would expend less energy burrowing in glass beads than in sand, leading to underestimations of burrowing cost.
We found that for most species in most materials, variability in velocity was fairly low, but combining species and materials to obtain a range of velocities showed that increased burrowing velocity correlated to increased cycle frequency rather than to increased cycle distance (Fig. 7). This pattern should be viewed with some caution, as behavioral differences among both species and materials exist. For crawling earthworms, however, velocity correlates more strongly with cycle distance than with frequency (Quillin, 1999). While we cannot discount the possibility that terrestrial and marine annelids exhibit different loco-motory behaviors, this difference could indicate that annelids use different strategies for crawling and burrowing, consistent with the different environmental forces for the two locomotory behaviors (Dorgan, 2010). Burrowing animals experience much greater anterior resistance to forward movement (e.g., work of fracture or plastic deformation for muds or sands, respectively) than frictional resistance for crawling (Dorgan, 2010). It seems plausible that higher anterior resistance might limit cycle distance in burrowing and result in different strategies to increase velocity for the two modes of locomotion. It would he interesting to compare crawling and burrowing within the same species to determine whether these different strategies to increase velocity depend on locomotory mode rather than simply reflecting species differences.
The morphological and behavioral differences among these three species are consistent with their habitats (Table 1).A wider anterior (N. dendritica), larger parapodia (N. dendri tice), and possibly also ventral papillae (O. Johnsoni) increase surface area and consequently friction, reducing backward slipping in sand. While the length of parapodia was not measured due to the quality of the video, observations of burrowing worms showed that N. dendritica had considerably longer parapodia compared to L. pugettensis and O. Johnsoni. Whether the ancestral orbiniid had a wider rounded anterior or a pointed anterior is unclear from the most recent phylogeny (Bleidorn et al., 2009), so we cannot determine whether a rounded anterior evolved in N. den-dritica to adapt to a sandy environment or whether a pointed anterior evolved in L. pugettensis and O. Johnsoni from an ancestor with a rounded anterior. Moreover, the shape of the anterior can change with ontogeny; juvenile L. pugettensis have a rounded anterior (Bleidorn et al., 2009). These differences, however, do seem to be functionally important. For mud burrowers like L. pugettensis, the narrow head can act as a wedge and extend the burrow via crack extension, and smaller worms are predicted by fracture mechanics theory to have blunter (or more rounded) antcriors (Che and Dorgan, 2010). Thetwisting behavior exhibited to varying extents by all three species has not, to our knowledge, been described in other polychaetes. Twisting was most common and periodic in L. pugettensis, and we suggest that this behavior increases dorsoventral forces on the material wall to extend the burrow by fracture. Perhaps 0.johnsoni's apparently aperiodic twisting helps it pack the burrow walls to prevent collapse. Although we collected N. dendritica and O. Johnsoni from the same sandy beach, N. dendritica is also found along the California coast in the rocky intertidal in algal holdfasts and debris around rocks, whereas O. Johnsoni is primarily found on sandy beaches (Blake and Ruff, 2007). This is consistent with our observations that O. Johnsoni appeared to exhibit a more specialized behavior for sands, expanding the thoracic region to pack the burrow walls, whereas N. dendritica showed a more intermediate burrowing strategy. L. pugettensis is found in a range of muds to sands (Blake and Ruff, 2007), but its burrowing behavior seems adapted to sediments with at least some amount of cohesion. Even intermediate muddy sands have been shown to crack in response to hydraulic pressure (Matsui et al.,2011), indicating that these intermediate sediments are cohesive enough for burrowing by fracture.
This study demonstrates how evolutionary divergence has led to fairly small but probably functionally important differences in morphologies and behaviors that are consistent with the constraints of these different physical environments. While all of the worms could burrow in both mud and sand, and thus could hypothetically inhabit both materials, their morphologies and behaviors appear to be well suited for their natural environments.
This project was funded by NSF OCE grant no. 1029160. We thank G. Rouse for helpful discussions and for the photos in Fig. I A, B. We also thank A. Chisholm for data on worm species locations. Nils Volkenborn provided helpful comments on the manuscript and compiled the supplementary video.
Blake, J. A., and R. E. Ruff. 2007. Polychaeta. Pp. 309-410 in The Light and Smith Manual. Intertidal invertebrates From Central California to Oregon, 1. T. Carlton, ed. University of California Press, Berkeley, CA.
Bleidorn, C., N. Hill. C. Erseus, and R. Tiedemann. 2009. On the role of character loss in orbiniid phylogeny (Annelida): molecules vs. morphology. Mol. Phylogenet. Evol. 52: 57-69.
Boudreau, B. P., C. Algar, B. D. Johnson, 1. Croudale, A. Reed, Y. Furukawa, K. M. Dorgan, P. A. Jumars, A. S. Grader, and B. S.
Gardiner. 2005. Bubble growth and rise in soft sediments. Geology 6: 517-520.
Che, J., and K. M. Dorgan. 2010. It's tough to be small: dependence of burrowing kinematics on body size. J. Exp. Biol. 213: 1241-1250.
Dorgan, K. M. 2010. Environmental constraints on the mechanics of crawling and burrowing using hydrostatic skeletons. Exp. Mech. 50: 1373-1381.
Dorgan, K. M., P. A. Jumars, B. D. Johnson, B. P. Boudreau, and E. Landis. 2005. Burrow extension by crack propagation. Nature 433: 475.
Dorgan, K. D., P. A. Jumars, B. D. Johnson, and B. P. Boudreau. 2006. Macrofaunal burrowing: The medium is the message. Oceanogr. Mar. Biol. 44: 85-121.
Dorgan, K. M., S. R. Arwade, and P. A. Jumars. 2007. Burrowing in muddy sediments by crack propagation: forces and kinematics. J. Exp. Biol. 210: 4198-4212.
Dorgan, K. M., S. Lefebvre, J. H. Stillman, and M. A. R. Koehl. 2011. Energetics of burrowing by the cirratulid polychacte Cirriformia moorei. J. Exp. Biol. 214: 2202-2214.
Duran, J. 2000. Sands. Powders, and Grains: An Introduction to the Physics of Granular Materials. Springer, New York.
Fauchald, K., and P. A. Jumars. 1979. The diet of worms: a study of polychacte feeding guilds. Oceanogr. Mar. Biol. 17: 193-284.
Geng, J., D. Howell, E. Longhi,R. Behringer, G.Reydellet, L. Vanel, E. Clement,and S. Luding.2001. Footprints in sand: the response of a granularmaterial to local perturbations. Phys. Rev. Lett. 87: 035506.
Goldenberg, C., and I. Goldhirsch. 2005. Friction enhances elasticity in granular solids. Nature 435: 188-191.
Herman, P., J. J. Middelburg, and C. Heip. 2001. Benthic community structure and sediment processes on an intertidal flat: results from the ECOFLAT project. Cont. Shelf Res. 21: 2055-2071.
Johnson, B. D., B. P. Boudreau, B. S. Gardiner, and R. Maass. 2002. Mechanical response of sediments to bubble growth. Mar. Geol. 187: 347-363.
Josephson, R. K., and K. W. Flessa. 1972. Cryolite: a medium for the study of burrowing aquatic organisms. Limnol. Oceanogr. 17: 134 - 135.
Law, C., K. M. Dorgan, and G. W. Rouse. 2014. Relating divergence in polychaete musculature to different burrowing behaviors: a study using Opheliidae (Annelida). J. Morphol. doi: 10.1002/jmor.20237.
Maeda, K., H. Sakai, A. Kondo, T. Yamaguchi. M. Fukuma, and E. Nukudani. 2010. Stress chain based micromechanics of sand with grain shape effect. Gramsl. Matter 12: 499-505.
Matsui, G. Y., N. Volkenborn, L. Polerecky, U. Henne, D. S. Wethey, C. R. Lovell, and S. A. Woodin. 2011. Mechanical imitation of bidirectional bioadeviction in aquatic sediments. Limnol. Oceanogr. Methods 9: 84-96.
Meijering, E., 0. Dzyubachyk, and I. Smal. 2012. Methods for cell and particle tracking. Methods Enzymol. 504: 183-200.
Meysman, F. J. R., J. J. Middleburg, and C. H. R. Help. 2006. Bioturbation: a fresh look at Darwin's last idea. Trends Ecol. Evol. 12: 688-695.
Murphy, E. A. K., and K. M. Dorgan. 2011. Burrow extension with a proboscis: mechanics of burrowing by the glycerid Hemipodus simplex. .1. E.zp. Biol. 214t 1017-1027.
Quillin, K. J. 1999. Kinematic scaling of locomotion by hydrostatic animals: ontogeny of peristaltic crawling by the earthworm Lumbricus terrestris. J. Exp. Biol. 202: 661-674.
Rouse, G. W., and F. Pleijel. 2001. Polychaetes. Oxford University Press, New York.
Rouse, G. W., and F. Pleijel. 2007. Annelida. Zootaxa 1668: 245-264. Rousset, V., F. Pleijel, G. W. Rouse, C. Erseus, and M. E. Siddall. 2007. A molecular phylogeny of annelids. Cladistics 23: 41-63.
Sanders, H. L. 1958. Benthic studies in Buzzard's Bay: animal-sediment relationships. Limnol. Oceanogr. 3: 245-258.
Struck, T. H., C. Paul, N. Hill, S. Hartmann, C. Hose, M. Kube, B. Lieb, A. Meyer, R. Tiedcmann, G. Purschke, and C. Blcidorn. 2011. Phy-logenomic analyses unravel annelid evolution. Nature 471: 95-98.
Thomas, P. A., and J. D. Bray. 1999. Capturing nonspherical shape of granular media with disk clusters. J. Geotech. Geoenviron. Eng. 125: 169-178.
Trucman, E. R. 1975. The Locomotion of Soft-Bodied Animals. American Elsevier Publishing. New York.
Winter, A. G., R. L. H. Deits, and A. E. Hosoi. 2012. Localized fluidization burrowing mechanics of Ensis directus. J. Exp. Biol. 215: 2072-2080.
Zrzavy, J., R. Pavel, P. Lubomir, and J. Janouskovec. 2009. Phylog-eny of Annelida (Lophotrochozoa): total evidence analysis of morphology and six genes. BMC Evol. Biol. 9: 189-193.
ALEX A. FRANCOEUR AND KELLY M. DORGAN (1),*
Scripps Institution of Oceanography, 9500 Gihnan Drive, La Jolla, California 92093
Received 16 December 2013; accepted 20 February 2014.
To whom correspondence should be addressed. E-mail: email@example.com
(1)Current address: Dauphin Island Sea Lab, 101 Bienville Blvd. Dauphin Island, AL 36528.
|Printer friendly Cite/link Email Feedback|
|Author:||Francoeur, Alex A.; Dorgan, Kelly M.|
|Publication:||The Biological Bulletin|
|Date:||Apr 1, 2014|
|Previous Article:||Field study of growth and calcification rates of three species of articulated coralline algae in British Columbia, Canada.|
|Next Article:||Melatonin production in the sea star echinaster brasiliensis (Echinodermata).|