ABSTRACT For this extraordinary postnova, based on 336 measures on 29 nights in August-October 1978, we find a peak-to-peak photometric variation in unfiltered light of 0.35 magnitude for the well-known 3.35-hour periodicity. Our period of 0.139617 ± 0.000005 day, obtained by a power-spectrum method, agrees with Patterson's determinations for both 1977 and 1978, which he found from the minima timings. In the variable light-curve shape and amplitude the period has clearly stabilized and undoubtedly reflects the binary orbital motion. A search for the long-surmised "double" period of 6.7 hours is described. Our mean light curve for the three-month interval shows no evidence, within quite small statistical fluctuations (Δm ≤ 0.03), for the double period. While a double-period process (conceivably the true orbital period) cannot be totally ruled out as yet, it seems more likely now that the apparent double light curves seen in limited time intervals can be attributed to random or quasi-random effects on a time scale of one to several days, e.g., changes in the shape of the envelope around the binary nucleus.