Colour and pattern change against visually heterogeneous backgrounds in the tree frog Hyla japonica

Abstract

Colour change in animals can be adaptive phenotypic plasticity in heterogeneous environments. Camouflage through background colour matching has been considered a primary force that drives the evolution of colour changing ability. However, the mechanism to which animals change their colour and patterns under visually heterogeneous backgrounds (i.e. consisting of more than one colour) has only been identified in limited taxa. Here, we investigated the colour change process of the Japanese tree frog (Hyla japonica) against patterned backgrounds and elucidated how the expression of dorsal patterns changes against various achromatic/chromatic backgrounds with/without patterns. Our main findings are i) frogs primarily responded to the achromatic differences in background, ii) their contrasting dorsal patterns were conditionally expressed dependent on the brightness of backgrounds, iii) against mixed coloured background, frogs adopted intermediate forms between two colours. Using predator (avian and snake) vision models, we determined that colour differences against different backgrounds yielded perceptible changes in dorsal colours. We also found substantial individual variation in colour changing ability and the levels of dorsal pattern expression between individuals. We discuss the possibility of correlational selection on colour changing ability and resting behaviour that maintains the high variation in colour changing ability within population.

Figure 1. Background colours that were used for BW (upper panels) and GB (lower left three panels) experiments.

Black (BK), black/white check-patterned (BW), grey (GY), white (WH), green (GR), green/brown check-patterned (GB), and brown (BR) background. The photo in the lower right side is an example of H. japonica with contrasting dorsal patterns against BW background.

Figure 2. The relationship between time passed since the initiation of each colour changing trial and the Euclidean colour distance (∆E) from the initial status in BW experiment (N = 48).

In all four types of backgrounds, frog colours changed rapidly within one hour and maintained similar levels throughout the remaining time. Symbols represent mean values and bars denote standard error of the mean.

Figure 3

(A) Brightness, (B) chroma, and (C) hue of dorsal colour of frogs against each achromatic backgrounds (N = 48). Symbols represent mean values and bars denote standard error of the mean.

Figure 4. Achromatic contrasts between reference colours and frogs’ dorsal colours against each background type.

Reference colours were black (BK) and white (WH). We derived just noticeable differences (JNDs; see methods) from (A) avian and (B) snake vision model. A frog may be detected by the predator when JNDs >1. Symbols represent mean values and bars denote standard error of the mean.

Figure 5. Proportion of patterned individuals against each achromatic background (A; N = 48) and the relationship between individual colour changing capacity and the presence of dorsal patterns against BK background (B; N = 72).

(A) Frogs exhibited dorsal patterns more often when they were against darker backgrounds rather than brighter backgrounds. (B) Frogs with higher colour changing capacity were more likely to exhibit dorsal patterns.

Figure 6. Colour change of frogs during night-time under darkness in BW experiments (N = 48).

Frogs retained substantial colour properties of the previous day in terms of both (A) brightness and (B) chroma. Symbols represent mean values and bars denote standard error of the mean.