Linking hunting weaponry to attack strategies in sailfish and striped marlin

Abstract

Linking morphological differences in foraging adaptations to prey choice and feeding strategies has provided major evolutionary insights across taxa. Here, we combine behavioural and morphological approaches to explore and compare the role of the rostrum (bill) and micro-teeth in the feeding behaviour of sailfish (Istiophorus platypterus) and striped marlin (Kajikia audax) when attacking schooling sardine prey. Behavioural results from high-speed videos showed that sailfish and striped marlin both regularly made rostrum contact with prey but displayed distinct strategies. Marlin used high-speed dashes, breaking schools apart, often contacting prey incidentally or tapping at isolated prey with their rostra; while sailfish used their rostra more frequently and tended to use a slower, less disruptive approach with more horizontal rostral slashes on cohesive prey schools. Capture success per attack was similar between species, but striped marlin had higher capture rates per minute. The rostra of both species are covered with micro-teeth, and micro-CT imaging showed that species did not differ in average micro-tooth length, but sailfish had a higher density of micro-teeth on the dorsal and ventral sides of their rostra and a higher amount of micro-teeth regrowth, suggesting a greater amount of rostrum use is associated with more investment in micro-teeth. Our analysis shows that the rostra of billfish are used in distinct ways and we discuss our results in the broader context of relationships between morphological and behavioural feeding adaptations across species.

Keywords: attack behaviour; billfish; feeding specialization; morphology; sailfish (Istiophorus platypterus); striped marlin (Kajikia audax).

Figure 1.

Examples of micro-CT overview images displaying the positions of all micro-teeth measured on the rostrum, coloured by micro-teeth type, for (a) one sailfish and (b) one striped marlin specimen. Green dots/arrows represent intact micro-teeth, pink dots/arrows represent broken micro-teeth and yellow dots/arrows represent re-growing micro-teeth in all images. The x-axis (h, mm) represents the distance from bill tip (0 on the far left) to the head (50 on the far right). The y-axis and purple lines show the angular position of the micro-teeth in degrees from the roll-axis of the rostrum. 0° is the dextral side, 90° is the dorsal side, 180° is the sinistral side and 270° is the ventral side. (c) 3D computer visualization with volume rendering of a random section of rostral surface displaying micro-teeth, (d) 3D surface of sailfish rostrum showing micro-teeth, (e) 3D surface of marlin rostrum showing micro-teeth, (f) sailfish rostrum cross-section (21.26 mm from tip) (scale bar, 500 µm), (g) marlin rostrum cross-section (21.26 mm from tip) (scale bar, 500 µm), (h) 3D perspective of the tip of sailfish rostrum, (i) slice of sailfish rostrum showing (yellow arrow) re-growing micro-tooth encroaching on broken dead micro-teeth, and (orange arrow) pulp of healthy tooth (scale bar, 200 µm), (j) micro-CT of a single intact micro-tooth showing enamel cap (red arrow) and organic pulp cavity (orange arrow) (scale bar, 100 µm), (k) micro-CT image of entire sailfish rostrum tip (scale bar, 2 mm). (Online version in colour.)

Figure 2.

Attack sequences for (a) sailfish and (b) striped marlin in the form of a Markov chain with state transitions. Only those transitions that were observed at least three times are shown. In order to make the figures comparable, the counts of the transitions for the striped marlin data were scaled, such that their sum equalled the sum of the transition counts for the sailfish data. States (approach, dash, bill use, prey contact, open, ingest, leave) are shown within ellipses and transitional probabilities are represented by the weighted arrows linking states. Photo credit: Rodrigo Friscione Wyssmann. (Online version in colour.)

Figure 3.

Rostrum use, and capture success for sailfish and striped marlin. (a) The likelihood of bill use during an attack sequence, (b) the likelihood the type of bill use was a slash, (c) the likelihood that bill was used on isolated prey, (d) the likelihood that the bill was used in a horizontal motion. (e) The likelihood that an attack sequence resulted in a successful capture, and (f) the mean capture rate per minute per group. (a–e) The marginal effects of the brms, with the dots showing the posterior median and the errors bars the 95% credibility intervals. (Online version in colour.)

Figure 4.

Micro-teeth analysis from micro-CT for sailfish (n = 6, blue) and striped marlin (n = 4, red) rostral tips (ranging from 0 to 55 mm). Shown are the effect of ‘species', ‘angle' and ‘distance from tip' on (a–c) mean tooth length, (d–f) fraction of broken teeth, (g–i) fraction of growing teeth, and (j–l) tooth density. Shown are the marginal effects of the brms, with the solid lines/dots showing the posterior median; and the shaded areas/error bars the 95% credibility intervals. (Online version in colour.)