In order to verify the consistency between the observed colors and the reflection spectrum of individual iridocytes, we targeted blue, green, yellow as well as orange-red bright cells on the shallow mantle of all specimens.

When the observed color of iridocytes was blue, the reflectance peak maximum was at ∼480 nm. In addition, the peaks of spectra curves of green, yellow, and orange-red iridescent cells were located at ∼570 nm, ∼590 nm, and ∼620 nm, respectively ( Figure 3 column 2). As expected, the peaks result in the observed colors of individual cells. In the visible spectrum, the blue range is 440 to 480 nm, the green range is 500 to 570 nm, the yellow range is 580 to 590 nm, and the orange-red range is 600 to 700 nm. Such structural coloration is an optical process produced by the submicroscopic structure of organisms, rather than related to pigmentation. That indicates that some changes in structure at nanoscale within iridocytes should be accountable for the various structural colors displayed in the mantle.

The TEM images in Figure 3 column 3 show the inner structure of iridocytes at the nanoscale. As mentioned in previous studies, most iridescent cells are occupied by stacked flattened lamellae, which are arranged in a uniform orientation (Kamishima, 1990; Griffiths et al., 1992; Kawaguti, 1966). Between the neighboring lamellae is the cytoplasm, which is enveloped with a membrane called cisterns. Electron dense lamellae with a high refractive index of 1.44 and the cisterns with a low refractive index of 1.35 are alternately arranged to form a periodic structure similar to the Bragg reflector. One pair of layers with high and low refractive index are considered as basic reflection units, which can enhance the reflection of a specific wavelength of light. Additionally, a quarter of the increased reflectivity wavelength is equal to the optical thickness (product of physical thickness and refractive index) of the layer pair.

The optical thickness of iridocytes of T. crocea is variable in a certain range. We measured the thickness of lamellae and cisterns of individual iridocytes to study the relationship between nanostructure and particular colors. The average recorded lamellar thickness was 76.84 ± 6.45 nm, ranging from 57.23 ± 4.97 nm to 115.53 ± 6.92 nm. The measurement is summarized in Table 1 .

The reflectance spectrum of the first model designed to simulate the optical characteristics corresponding to the cell structures of blue, green, and yellow iridocytes is shown in Figure 4 . After the simulation was completed, the origin wavelength excitation modeled from 300 to 800 nm showed varying degrees of reflection. Some wavelengths of light were strongly enhanced, while a great part of wavelengths were eliminated by interference. The blue, green, and yellow curves represent the wavelength reflection of the iridocytes structure of the corresponding color. The simulated peak wavelength varies in the visible range, from blue band to yellow band, including three isolated and intensive one-peak curves. As expected, the peak wavelength of blue iridocytes is approximately 460 nm, and the reflection value in the blue band in the visible spectrum is 440 to 480 nm. The green curve and yellow curve move to the right relative to the blue curve, with peaks of 540 nm and 580 nm, respectively, corresponding to the green and yellow bands. Figure 4 reveals that with the increase of the thickness and spacing of lamellae, the wavelength of the reflection peak increases, and the corresponding light band in the visible spectrum shifts toward red. To further investigate how the structure determines the color of iridocytes, it is necessary to quantify these two parameters, namely, the thickness and spacing of layers.

The result of the second is shown in Figure 5 . The simulation was conducted by increasing h₁ every 10 nm with changing h₂ within a specified range to obtain the intensity of the reflection. When h₁ is fixed and h₂ increases from 40 nm to 110 nm, as seen from Figures 5A–G , the reflected wavelength gradually increases, shifts from ultraviolet to blue, and then towards red. In Figures 5A–H , when h₂ remains unchanged and h₁ increases, the wavelengths increase and the reflected wave bands become wider at the same time. It is worth noting that when h₁ reaches 120 nm and h₂ exceeds 90 nm, two bright bands appear in the reflection distribution ( Figure 5H ), with strong reflection in both visible wavelength and ultraviolet regions. Similar cases also appear in Figure 5I , h₁ = 130 nm and h₂ above 80 nm; Figure 5J , h₁ = 140 nm, h₂ above 70 nm; Figure 5K , h₁ = 150 nm, h₂ above 60 nm. More importantly, with the increase of thickness, iridocytes take successively multiple colorations of green, yellow, or even orange-red. Surprisingly, when the sum of thickness and spacing of lamellae reach above 210 nm, iridocytes seem to produce both visible and ultraviolet radiation, which has not been found in previous studies ( Figures 5I–K ).

Previous studies have preliminarily shown that the structural colors are determined by nanostructure (Ghoshal et al., 2016a). In our study, we found that an optical process is of great importance in iridocytes’ colors generated by the lamellae structure. That indicates that some changes in structure at the nanoscale within iridocytes should be accountable for the various structural colors displayed in the mantle. Our FDTD model is designed for the simulation of iridocytes with alternating high and low refractive index layers observed in T. crocea and applies to cells with similar structures. Furthermore, the ultrastructure structures of iridescent cells in T. maxima and T. gigas are consistent with that of iridocytes in T. crocea. Therefore, our model holds true in T. maxima and T. gigas. The results of the FDTD simulations show that there is a clear corresponding relationship between coloration and the internal Bragg stacks, which is associated with thicknesses and spacing of lamellae. Since the lamellae are thought to be a stable substance with constant thickness in iridocytes, in real situations, the variation of colors is more likely to be caused by the change of spacing. Similar to the lamellae in the iridocytes of giant clams, squid also has an unusual structure called reflectins that create thin-film interference, resulting in the reflection of certain incident light (Crookes et al., 2004; Kramer et al., 2007). The reflections in squid are connected to the extracellular environment. When the water flows between the reflectors and the extracellular space, the spacing of lamellae becomes larger or smaller, resulting in color variation of squid (DeMartini et al., 2013). We speculated that this phenomenon might also exist in giant clams, and studied the theoretical possibility of creating FDTD simulation of iridocytes in T. crocea. We also found that the structural colors caused by the thickness and spacing of laminae drive the reflected light of iridocytes to vary in the full spectrum as well as the ultraviolet region. However, different species or strains with typical coloration and patterns usually maintain their characteristics. Therefore, the Bragg structure of individual iridocytes may also stay in a stable range.

In addition, there is another interesting phenomenon in the FDTD simulation results: when the lamellar structure meets certain conditions, iridocytes will generate ultraviolet radiation, which is not useful for giant clams. In general, low dose ultraviolet would benefit organisms with sterilization, while high energy strong ultraviolet radiation (280-400 nm) often imposes adverse effects, such as DNA damage, oxidative stress, and inhibition of photosynthesis (Lesser, 1996; van de Poll et al., 2001; Llabres et al., 2013). To protect the giant clam itself, there are countless eyes and photosensitivity cells along the margin of the mantle, which can sense light (Wilkens, 1984; Land, 2003). This visual system is also sensitive to ultraviolet radiation, thus enabling the giant clams to respond to a dynamic environment and to meet the requirements of clam symbionts (Rossbach et al., 2020a). Many marine bivalves can rapidly adapt to various environmental conditions. Marine bivalves have been found to actively modify the calcification physiology in response to an acidifying ocean (Zhao et al., 2020). In addition, pearl oysters can respond plastically to marine heatwaves with a series of enzymes activity enhanced (He et al., 2021). Previous studies have shown that giant clams can avoid high-intensity ultraviolet damage in some special ways. Iridocytes in the mantle are thought to eliminate part of ultraviolet radiation by performing non-photosynthetic backscattering (Ghoshal et al., 2016a). Some studies believe that the lamellae stacks of the iridocytes, guanine, absorb ultraviolet radiation to protect the giant clams and promote zooxanthellae photosynthesis by continuously emitting longer wavelengths of light (Rossbach et al., 2020b). The symbiotic microalgae harbored by giant clams, zooxanthellae, can absorb certain wavelengths of light for performing photosynthesis. Abundant zooxanthellae live in the mantle of giant clams and form the tertiary zooxanthellae tube (Norton and Jones, 1992). The iridescent cells are thought to be enhancers of zooxanthellae photosynthesis, improving the light-harvesting efficiency and solar energy conversion. More specifically, iridocytes over the surface of the mantle redeliver light to algal pillars with controlled doses (Holt et al., 2014) and selectively use some particular bands of light, resulting in iridescent colors. Therefore, the ultrastructure of iridocytes is closely related to the changes in the photic environment in giant clams, affecting the photosynthetic efficiency of symbiotic zooxanthellae. In addition, this may provide the driving force for the distribution of different clades of symbiosis that preferentially utilize different parts of the solar spectrum for photosynthesis. In addition, a special substance known as mycosporine-like amino acids (MAAs) was found in the mantle of giant clams, which plays the role of ultraviolet-resistance (Ishikura et al., 1997). The factors mentioned above provide photoprotection and promote the efficient use of sunlight energy. In this study, the simulation results show that iridocytes may also have the potential to produce minute doses of ultraviolet radiation, which means that ultraviolet radiation may impose both beneficial and adverse effects on the giant clams, depending on the ultraviolet radiation dose they receive or produce.