The result is also quite insensitive to light intensity. If the sunlight is attenuated without spectral change, the bandgap shifts to a shorter wavelength, but the absorptance spectra at higher costs Wnt inhibitor remain essentially unaltered, as shown by the dashed lines in Fig. 3 calculated for 1% of full sunlight. They do shift to shorter wavelengths if attenuation is carried out using a (smooth) black-body irradiance spectrum, in accordance with the findings of Björn (1976), but the irregular shape of the actual solar spectrum

at sealevel keeps the optimal absorption bands at high costs fixed in the same position. The QY absorption bands of chlorophyll a and b cover the spectral range between the 687 and 628 nm absorption bands of atmospheric O2, and are separated by the 656 nm H-α absorption line in the solar spectrum. When these O2 absorption bands were removed from the AM 1.5 spectrum, using the local shape of the AM 0 PXD101 spectrum with a slope correction, the optimized

absorptance band at high cost was still at the Chl a position, but jumped to the Chl b position when optimized at 1% of the light intensity. In order to determine if the similarity between real and predicted spectra in Fig. 4 is merely a coincidence, we applied Torin 2 in vivo the same analysis to one of the “colorful spectral niches” at the bottom of the photic zone described by Stomp et al. (2007). Figure 5 shows the solar irradiance under 5 cm of water with a high concentration of organic matter. At the same relative cost that yielded a good approximation of the red band of photosynthesis in non-attenuated sunlight, optimization for growth power in this spectral niche yields an absorptance spectrum that resembles the QY absorption of bacteriochlorophyll A in purple non-sulfur bacteria (Fig. 6). The lower and upper bounds of the spectral range depend on the arbitrary choice of water depth and organic matter concentration. The fact that the deep trough around 820 nm is reproduced by the

effect of a minor atmospheric H2O absorption band Methane monooxygenase on the optimization, however, does provide independent evidence for the validity of the analysis presented here. Fig. 5 The transmitted power spectra of Fig. 1 calculated for the irradiance in a muddy pool. To select the spectral range absorbed by bacteriochlorophyll A, the solar irradiance was attenuated by 5 cm water (Hale and Query 1973) with a “gilvin and tripton” attenuation coefficient K GT(440) = 11 cm−1 as described by Stomp et al. (2007). The same relative cost values as in Fig. 1 were used Fig. 6 The absorptance spectra of Fig. 4 calculated for the irradiance spectrum selected in Fig. 5. Growth power optimized absorptance spectra for the same relative cost values as in Figs. 3 and 4.