Frequency-dependent Batesian mimicry maintains colour polymorphism in a sea snake population

Abstract

Evolutionary theory suggests that polymorphic traits can be maintained within a single population only under specific conditions, such as negative frequency-dependent selection or heterozygote advantage. Non-venomous turtle-headed sea snakes (Emydocephalus annulatus) living in shallow bays near Noumea in New Caledonia exhibit three colour morphs: black, black-and-white banded, and an intermediate (grey-banded) morph that darkens with age. We recorded morph frequencies during 18 consecutive years of surveys, and found that the numbers of recruits (neonates plus immigrants) belonging to each morph increased in years when that morph was unusually rare in the population, and decreased when that morph was unusually common. Thus, morph frequencies are maintained by negative frequency-dependent selection. We interpret the situation as Batesian mimicry of highly venomous sea snakes (Aipysurus, Hydrophis, Laticauda) that occur in the same bays, and range in colour from black-and-white banded to grey-banded. Consistent with the idea that mimicry may protect snakes from attack by large fish and sea eagles, behavioural studies have shown that smaller fish species in these bays flee from banded snakes but attack black individuals. As predicted by theory, mimetic (banded) morphs are less common than the cryptically-coloured melanic morph.

The banded colour patterns of sea snakes discourage attack by predatory fishes, enabling Batesian mimicry by harmless species

“…This species also is distinctive in exhibiting colour polymorphism: only around 20% of individuals exhibit bands (see above). Our long-term (20-year) mark–recapture studies on these sea snakes [ 17 , 30 ] show that morph frequencies have remained stable. That is, a year with a high proportion of melanic individuals is followed by recruitment of a cohort containing a high proportion of banded individuals, and vice versa for the other colour morphs.…”

“…That is, a year with a high proportion of melanic individuals is followed by recruitment of a cohort containing a high proportion of banded individuals, and vice versa for the other colour morphs. Thus, the relative frequencies of the three morphs appear to be maintained by a selective advantage that accrues to whichever morph is less common [ 17 ].…”

“…Under this hypothesis, the relative frequencies of mimics are stable through time because the advantage to mimicry decreases when banding is no longer strongly associated with toxicity. If the banded morphs of E. annulatus become common, the deterrent value falls because predatory attacks to banded snakes are less likely to be followed by envenomation of the predator [ 12 , 13 , 17 ]. Empirically testing the frequency-dependence of any advantage to a specific colour morph is logistically challenging, and would require a system with spatial or temporal variation in morph frequencies.…”

“…In shallow bays around the city of Noumea in the IndoPacific archipelago of New Caledonia, turtle-headed sea snakes ( Emydocephalus annulatus ) exhibit three colour morphs [ 17 ] ( figure 1 ). About 80% of individuals are jet-black (melanic) throughout life, about 10% are brightly banded in black-and-white throughout life, and the other 10% are born with indistinct (grey) bands on a black background but transform to melanism as they age [ 17 ]. The banded morphs resemble sympatric deadly sea snakes [ 17 ].…”

“…About 80% of individuals are jet-black (melanic) throughout life, about 10% are brightly banded in black-and-white throughout life, and the other 10% are born with indistinct (grey) bands on a black background but transform to melanism as they age [ 17 ]. The banded morphs resemble sympatric deadly sea snakes [ 17 ]. If this situation arose via mimicry, we would predict that predators avoid banded snakes but attack black ones.…”