Fig 1.
Field observation of postprandial drinking behaviour of mudskippers.
(A) Example photographs showing terrestrial eating (upper) and migration to water (lower). Terrestrial eating of mudskippers is clearly observed because the fish put its mouth on the mudflat following bending of the body axis. Migration to water areas where mudskippers fully bathed their mouths was counted as an index of a desire to drink. (B) Correlation between frequencies of eating bouts and those of migration into water (n = 20) in a 5-min observation. (C) The behavioural transition between two modes: terrestrial eating and migration into water. Numbers represent transition probabilities calculated from bouts of terrestrial eating (n = 46) and migration into water (n = 106) from 20 fish. The red arrow shows postprandial drinking. (D) Measured and ‘presumed’ latency of migration into a water area after terrestrial eating. The calculation method is illustrated in S1 Fig. *P < 0.01 by paired t test. Data are shown as means ± standard error of the mean (SEM) and presented in S1 Table. Paired data from individual fish are shown by dotted lines.
Fig 2.
Schematic diagram of the experimental tank used to observe postprandial drinking of mudskippers.
10 ppt seawater is close to the natural environmental salinity and almost identical to the osmolality of body fluids, and thus was chosen for subsequent experiments. Mudskippers were transferred to the tank without food. After 60 min, food was placed on a land area.
Fig 3.
Postprandial drinking behaviour in aquaria.
(A) Amphibious behaviour before feeding and while eating. The frequency of migration and period of time in water were measured for 30 min. Paired data from individual fish are shown (red lines). A paired Wilcoxon signed-rank test was used for statistical analysis. † P < 0.05 vs. controls before feeding. Raw data are presented in S2 Table. (B) Effects of eating on time-course changes in amphibious behaviour. Data from fed fish (red lines) and unfed fish (black lines) are expressed as mean ± SEM with individual data. The parameters were measured at 5 (0–5), 10 (5–10), 15 (10–15), 20 (15–20), 25 (20–25), and 30 (25–30) min after onset of eating, as well at 0 (0–5) min (before feeding). Two-way repeated measures ANOVA and a Steel post-hoc test were used for statistical analysis. *P < 0.05, *** P < 0.001 vs. unfed controls. †P < 0.05, ††P < 0.01 vs. controls before feeding. (C) Effect of eating on drinking rate. An unpaired Wilcoxon signed-rank test was used for statistical analysis. **P < 0.005 vs. unfed controls. Individual data are plotted with the mean ± SEM and presented also in S3 Table.
Fig 4.
Rolling behaviour and pectoral-fin movements in dry land.
(A) Photographs showing movements of a pectoral fin (broken lines) with schematic drawings. The pectoral fins touch wet ground (left). The fin sequentially moves forward (middle) and then backward (right). A movie showing movements on dry land is provided in S1 Movie. (B) Schematic diagram of the experimental tank used to observe rolling behaviour and pectoral-fin movements in controlled humidity. A wet hygroscopic cloth was placed to allow each mudskipper to moisten the skin. (C, D) Effects of drying on frequency of rolling behaviour (C) and pectoral-fin movements (D) after 5 (0–5), 10 (5–10), 15 (10–15), 20 (15–20), 25 (20–25), and 30 (25–30) min. Data from fish in <40% (low) humidity (red lines) and fish in >80% (high) humidity (black lines) are expressed as mean ± SEM with individual data. *P < 0.05, **P < 0.01 by unpaired Wilcoxson-rank test, following a Friedman test. (E) Laterality in frequencies of rolling behaviour and pectoral-fin movements. Correlations of percentages of right-side rolling and right fin movements are shown. Black and red circles indicate data for fish in high and low humidity, respectively. R2 values were 0.37 using all data (upper panel, n = 10) and 0.64 using data from fish in low humidity (lower panel, n = 6).
Fig 5.
Effects of environmental salinities on rolling and fin movements.
(A) Rolling behaviour of mudskippers after acclimation to freshwater and 10, 20, and 30 ppt seawater. There were no significant differences by Kruskal-Wallis test (n = 5–7). (B) Pectoral-fin movements of mudskippers after acclimation to freshwater and 10, 20, and 30 ppt seawater. *P < 0.05 with Kruskal-Wallis followed by Steel-Dwass post-hoc test. Individual data are plotted with the mean ± SEM.
Fig 6.
A conceptual diagram of implications of osmoregulatory behaviours in mudskipper fish.
In semi-terrestrial mudskippers, peripheral sensations such as a dry mouth following eating and cutaneous drying induces postprandial drinking and rolling behaviour with pectoral-fin movements, respectively. To prevent systemic dehydration, it is likely that behaviours are triggered by anticipatory motivations. In addition, these behaviours may play important roles in transport of terrestrial prey in the digestive tract, cutaneous respiration, and salt secretion. Such multifunctional behaviours may allow mudskippers to survive in terrestrial habitats.