On the input side, we could identify neurons connecting one of the small subunits of the AOTu with the lateral triangle of the LALs (TuLAL1; n = 10; Figure 3F, Table 1). On the output side,

a characteristically shaped neuron was identified that connected large parts of the LAL with regions of the unstructured protocerebrum, located lateral and dorsal to the monarch central body (LAL-PC-neuron; n = 1; Figure 3G; Figure S1G). We next used intracellular recordings Neratinib order to examine the response properties of CC monarch neurons in the context of neural integration of the major skylight cues (polarized and unpolarized light stimuli). There were three considerations for evaluating these recordings. First, which skylight cues are actually processed by the monarch central brain? Second, how do monarchs resolve the directional ambiguity of skylight E-vectors ( Figure 1A)—that

is, do they use spectral gradients for distinguishing the solar and antisolar hemisphere, as suggested for locusts? The third issue was learn more defining how polarized and unpolarized light responses are integrated to ensure that E-vector tuning actually provides an accurate reflection of the solar azimuth over the course of the day ( Figures 1A and 1B). For recordings, a migratory butterfly was mounted in the recording setup. Two types of visual stimuli (linear polarized light and unpolarized light spots) were applied during experimentation (Figure 1C). Prior to recordings, all migrants were housed in 11 hr of light

and 13 hr of darkness, simulating outdoor lighting conditions at capture; physiological experiments were centered around Zeitgeber time (ZT) 5, which was 5 hr after lights-on, so that substantial variation in time-of-day of recording would not confound the results. Nonmigratory monarchs were used for initial recordings and some control experiments. When presented Sitaxentan with zenith-positioned polarized light in the UV range (365 nm), 33 neurons of the monarch brain responded with significant E-vector-dependent modulations of their spike frequency, as revealed by circular statistics (p < 0.05; Figure 4). Polarized UV light was used because the monarch DRA ommatidia only express a UV opsin ( Sauman et al., 2005) and have been shown to be maximally sensitive to wavelengths below 380 nm ( Stalleicken et al., 2006). Of the E-vector-responsive cells, 19 could be identified anatomically postrecording. All of these cells were components of the proposed polarization vision network described above. We identified seven TuLAL1 cells, six TL-type cells, two CL1 cells, an individual TB1 cell, and three CPU1 neurons; hence E-vector-dependent responses were present from the input stage of the polarization vision network (TuLAL1 cells) to the output stage of the CC (CPU1 cells).