Interaction among sea urchins in response to food cues

Interaction among sea urchins remains largely uninvestigated, although the aggregation of sea urchins is common. In the present study, 1, 15 and 30 sea urchins Strongylocentrotus intermedius (11.06 ± 0.99 mm in test diameter) were placed in a 1 m2 circular tank, respectively. Movement behaviors were recorded for 12 min to investigate potential interactions among sea urchins. After the 12-min control period, we added food cues into the tank and recorded the changes in sea urchins’ behaviors. For the first time, we here quantified the interactions among sea urchins in laboratory and found that the interactions varied with food cues and with different densities. The sea urchins dispersed in random directions after being released. There was no significant difference in the movement speed and the displacement of sea urchins among the three density groups (1, 15 and 30 ind/m2). The interaction occurred when sea urchins randomly contacted with the conspecifics and slowed down the movement speed. The speed of sea urchins after physical contacts decreased by an average of 40% in the density of 15 ind/m2 and 17% in the density of 30 ind/m2. This interaction resulted in significantly higher randomness in the movement direction and lower movement linearity in 15 and 30 ind/m2 than in 1 ind/m2. After the introduction of food cues, the movement speed, displacement and dispersal distance of sea urchin groups decreased significantly in all the three densities. The dispersal distance and expansion speed of sea urchins were significantly lower in 30 ind/m2 than those in 15 ind/m2. The present study indicates that the interaction among sea urchins limits the movement of individual sea urchin and provides valuable information into how large groups of sea urchins are stable in places where food is plentiful.

www.nature.com/scientificreports/ S. intermedius (11.06 ± 0.99 mm in test diameter) in a 1 m 2 circular tank for 12 min after the release of food cues into the tank. Based on our previous records of speed of movement of small S. intermedius (0.34 ± 0.10 mm/s) 26 , 1 m 2 circular tank and 12-min experiment period allow sea urchins to move freely and allow us to observe possible interactions among sea urchins. Before food cues experiments, movement behavior of small S. intermedius was recorded for 12 min as the control. The main purposes of the present study are to investigate: (1) whether sea urchins interact with conspecifics; (2) whether interactions affect the behaviors of sea urchins at different densities; (3) whether food cues affect the interaction among sea urchins at different densities.

Results
Individual behaviors and interactions. There was no significant difference in the movement speed and displacement during the 12 min of observations among the three densities of sea urchins (0.51 ± 0.04 mm/s and 291.60 ± 33.53 mm in 1 ind/m 2 ; 0.44 ± 0.02 mm/s and 230.45 ± 11.67 mm in 15 ind/m 2 ; 0.46 ± 0.04 mm/s and 245.72 ± 21.40 mm in 30 ind/m 2 , Fig. 1A,B; Supplementary Tables S1 and 2). By calculating the distance between the sea urchin and the center of the tank after 12 min ( CD = [x i − x tank ] 2 + y i − y tank 2 × k/n , (x i , y i ) is the coordinate of sea urchin i at the end of the period, (x tank , y tank ) is the center coordinates of the tank, k is the scale of the picture, n is the number of sea urchins in different density group), the mean centrifugal distance was significantly different among the three density groups. Individual sea urchins (1 ind/m 2 ) moved significantly farther outward from the center than those in 15 Table S3). By comparing the distance and displacement of the sea urchins, we obtained the linearity l of free movement of sea urchins in 12 min ( l i = displacement i distance i . Displacement i is the movement displacement of sea urchin i after 12 min. Distance i is the total movement length of sea urchin i within 12 min). The motion linearity of sea urchins in 15 ind/m 2 was significantly lower than that of individual sea urchin in 1 ind/m 2 (0.79 ± 0.04 in 1 ind/m 2 ; 0.69 ± 0.02 in 15 ind/m 2 ; 0.70 ± 0.02 in 30 ind/m 2 , Fig. 1F; Supplementary Tables S1 and 2). R is a measure of the randomness of movement, where 0 is completely random movement and 1 is completely directional movement ( R =   Table S1 and 4). Dispersal distance r is the movement distance of each sea urchin from its initial position  Table S3). Expansion speed v e (mm/min) is the average speed of all sea urchins in the group from the  Table S5).

Discussion
In this study, interaction occurred when sea urchins randomly contacted conspecifics. This physical contact had negative effects on the movement of sea urchins, in which the speed of sea urchins dropped by about 40% in 15 ind/m 2 and 17% in 30 ind/m 2 . This explains the field phenomenon that sea urchins stop for a few seconds after contact with a conspecific 12 . Due to the interaction among sea urchins, the motion linearity of sea urchins in groups was significantly lower than that of individual sea urchins, while the randomness of movement direction was significantly higher. After being released, the sea urchins in the group diverged in all directions. Moving away from each other would reduce the negative effect of interaction. This is consistent with the field observation that the dispersal among sea urchins reduces the competition within sea urchin groups 27  www.nature.com/scientificreports/   25 . Rapid dispersal of S. intermedius is important to prevent the formation of high-density aggregations and the negative effects of increased stocking density on small S. intermedius 28,29 . The present study found that the high-density group formed at the beginning prevented the dispersion of sea urchins within the group. The explanation is that interactions within the group negatively affect the movement behaviors of sea urchin (e.g. movement speed). Thus, we suggest that high density is not appropriate during the reseeding of S. intermedius because of the negative interactions among sea urchins. After the release of the food cues, movement speed of sea urchins significantly slowed down in all the three density groups. This behavioral strategy increases the duration of sea urchins in food patches 8 . In this experiment, sea urchins did not form significant aggregations under the condition of evenly distributed food. This is different from the results of food-induced sea urchin aggregation reported in the field 2,13,30 . This difference indicates that the aggregation of sea urchins highly depends on condition of the food sources. The negative effects of sea urchin interactions on movement consistently existed with the exposure to food cues. Further, we found that dense sea urchin groups spread out significantly less. The low mobility of sea urchins was thought to be a factor Behavioral changes (mean ± SEM) of sea urchins exposed to the food cues when they randomly contacted conspecifics in different density groups. Movement linearity, mean result length R and dispersal distance of sea urchins exposed to the food cues in 1, 15 and 30 ind/m 2 (A-C). Speed of sea urchins before, during and after the contact in 15 ind/m 2 when exposed to the food cues (D, left) and speed per 5 s (D, right). Speed of sea urchins before, during and after the contact in 30 ind/m 2 when exposed to the food cues (E, left) and speed per 5 s (E, right). Expansion speed of sea urchins exposed to the food cues in 15 and 30 ind/m 2 (F-H). www.nature.com/scientificreports/ in the stable existence of a dense sea urchin on grazing front 12 . The present study suggests that low mobility of a grazing front is not only due to the decrease in sea urchins' speed in response to food cues, but the constrained expansion caused by the interaction among the sea urchins. The present study quantified the interaction among sea urchins and found negative impacts of sea urchins on the movement of conspecifics. This effect slowed the outward expansion of the sea urchin group and increased with the increasing density. Food cues caused sea urchins to slow down. Large groups spread more slowly in response to food cues due to the interaction among sea urchins. The present study suggests that high density is not appropriate in reseeding because of the adverse interaction among sea urchins. Before the experiments, the sea urchins were maintained at 12 ± 0.5 °C in a 300 L tank in the laboratory and were fed ad libitum fresh brown algae Gracilaria lemaneiformis. All applicable international, national, and/or institutional guidelines for the care and use of animals were followed by us.

Methods
Interactions among sea urchins. To investigate whether sea urchins interact with conspecifics, we measured the behaviors of 1, 15 and 30 sea urchins in a 1 m 2 circular tank, respectively. Fresh seawater was filled into the experimental device in a depth of 3 cm at the same temperature as the sea urchin's holding tank (12 ± 0.5 °C). Sea urchins were gently placed in the center of the device. There was no contact among sea urchins in groups at the beginning of the experiment (Fig. 4). The behaviors of individuals in the three groups were recorded for 12 min using a video camera (Legria HF20; Canon, Tokyo, Japan) without external food cues. Experiments were repeated eight times using different sea urchins for all the three densities (n = 8).
Behaviors and interactions among sea urchins exposed to food cue. After the 12 min of observation, we added 100 mL of the food cues over a period of two minutes and then recorded the response of the sea urchins to the food cues for another 12 min. The whole experiment lasted 26 min (Fig. 4). In the experiment, the food cues were produced by adding 20 g dry powder of kelp (S. japonica) to 100 mL fresh seawater and then filtering it three times with a bolting-silk net (44 µm in mesh size). The experiment was conducted at a low light intensity (8 lx). We repeated the experiment eight times using different sea urchins and changed the seawater for all three densities (n = 8).
Behaviors. Images from the 26-min video were taken every 5 s. A total of 313 images were combined into a stack using the software ImageJ (1.52 s version). We designated two periods of images in the stack: the period without food cues from 1 to 145 images, the food cue period from 169 to 313 images. For each video, we used the plugin manual tracking in ImageJ (1.52 s version) to extract the coordinates of each sea urchin.
We calculated the movement speed (v), and displacement (d) of the sea urchins as follows: www.nature.com/scientificreports/ where (x i (t), y i (t)) is the coordinates of sea urchin i at image t, (x p , y p ) is the initial position of each period of the sea urchin, k is the scale of the picture. During the experiment, sea urchins were considered in contact when they had physical contact with a conspecific (the touch of spines). Movement speeds were recorded from the beginning of contact to the separation of the sea urchins (no physical contact). The movement speeds in the 30 s before and after the contact process were recorded as the behavior before and after contact.
To measure the random level of movement, we measured the mean result length R of movement per minute for 12 min. A value of 0 indicates absolute random motion and a value of 1 indicates absolute directional movement 31 .
where θ t represents the turn angle at time t.
We measured the linearity l in each period using the ratio of displacement (d) to distance after 12 min 32 .
We calculated the distance of the sea urchins from their initial position at the end of the experiment. We used the average of this distance within the sea urchin group to indicate the spread r of different density groups.
where (x i , y i ) is the coordinate of sea urchin i at the end of the period, (x p , y p ) is the initial position of each period, (x 0 , y 0 ) is the coordinate at the beginning of the whole experiment, k is the scale of the picture, n is the number of sea urchins in different density group (15 ind/m 2 group: n = 15, 30 ind/m 2 group: n = 30).
The expansion speed of the sea urchin groups was compared with the distance of the individual from the center of the group. The center coordinates of the group (x c , y c ) are the mean of all the sea urchin coordinates in the group at the end of each period. Expansion speed v e (mm/min) is the average speed from group center for all sea urchins in each group.
where (x i , y i ) is the coordinate of sea urchin i at the end of the period, (x p , y p ) is the initial position of each period, (x c , y c ) / (x c0 , y c0 ) is the mean of all the sea urchin coordinates in the group at the end/beginning of each period. Statistical analysis. The data were tested for homogeneity of variance and normal distribution before all statistical analyses using the Levene test and Kolmogorov-Smirnov test, respectively.
In the period without food cues, movement speed, displacement, centrifugal distance, linearity and mean result length R of sea urchins among the three groups were analyzed using one-way ANOVA (1 ind/m 2 group: n = 8; 15 ind/m 2 group: n = 8; 30 ind/m 2 group: n = 8). Mann-Whitney U test was used if the data did not meet the normal distribution and/or variance. One-way repeated measures ANOVA was used to compare movement speeds of the sea urchin before, during, and after contact.
In the food cues experiments, paired sample T test was used to compare the movement speeds, displacement and expansion speed of sea urchins in response to the food cues (1 ind/m 2 group: n = 1 × 8 = 8; 15 ind/m 2 group: n = 15 × 8 = 120; 30 ind/m 2 group: n = 30 × 8 = 240). Wilcox signed-rank test was used for the data that did not satisfy the normal distribution. Movement linearity, mean result length R and dispersal distance among the three groups were analyzed using One-way ANOVA. Mann-Whitney U test was used when the data did not meet the normal distribution and/or variance. The independent sample T test was used to compare the speed and expansion speed between 15 and 30 ind/m 2 group after the food cues was released.
All data analyses were performed using SPSS 25.0 statistical software. A probability level of P < 0.05 was considered as significant.
Ethical approval. All applicable international, national, and/or institutional guidelines for the care and use of animals were followed by the authors.

Data availability
All data generated or analyzed during this study are included in this published article (and its Supplementary Information files).