All of the Tg(gng8:GAL4,UAS:EGFP-Δclocka) fish used in the following experiments were first screened for GFP fluorescence at 3 dpf using an Olympus MVX10 or Zeiss V16 stereo fluorescence microscope. Since the expression of EGFP-ΔCLK in the habenula was mosaic, only individuals with expression in the majority of habenula cells were selected and used for further experiments as ΔCLK-positive fish. To maximize expression, fish were generated by in-crossing. A consequence of this, however, is that fish with zero expression were rare. Thus, in addition to individuals with zero expression, larvae exhibiting a minimal number of randomly labelled cells (less than 15 cells per fish) were used as the control siblings (Fig. S1 ). Despite this, the size of the control group was smaller than the ΔCLK group.
Following the behavioural experiments, larvae were immobilized in tricaine methanesulfonate (MS-222; final concentration of 0.01 mg/mL; Sigma), mounted in 2% LMA in E3 medium and the habenula-specific expression of EGFP was assessed using a Zeiss LSM800 confocal microscope and 40X water dipping objective.
Following the behavioural experiments, larvae were immobilized in tricaine methanesulfonate (MS-222; final concentration of 0.01 mg/mL; Sigma), mounted in 2% LMA in E3 medium and the habenula-specific expression of EGFP was assessed using a Zeiss LSM800 confocal microscope and 40X water dipping objective.