The bolometric light curve of iPTF14hls (
a) deduced from the blackbody fits shows a late-time decline rate that is slower than the radioactive decay of 56Co (black), but consistent with both delayed accretion power (blue; t 0 is the onset of accretion at the last peak which could represent a final fallback event) and magnetar spindown power (red; t 0 is the formation time of the magnetar, P 0 is the initial spin period and B is the magnetic field in this simple analytic model). The magnetar model, however, is not consistent with the luminosity during the first 100 days, as implied by the P48, CSS and Gaia observations ( b), unless the early-time magnetar emission is substantially adiabatically degraded. TNT photometry of iPTF14hls and publicly available CSS photometry (retrieved from the CSS website) and Gaia photometry (retrieved from the Gaia Alerts website) not presented in Fig. 1 are shown in b. Data from the P48 (dashed lines) and the LCO 1-m telescope (solid lines) presented in Fig. 1 are shown for comparison. Photometric points from the same day, instrument and filter are averaged for clarity. The B - V ( c) and V - I/ i ( d) colour evolution of iPTF14hls from the LCO 1-m data (filled squares) differs from that of the normal type II-P SN 1999em (empty circles) , even when contracting the iPTF14hls data by a factor of 10 in time (empty squares) to compensate for the slow evolution observed in its spectra compared to that of normal type II-P supernovae. All error bars, when available, denote 1 22 σ uncertainties.