An article entitled "Three New Variables" appears in Cragg, T, PASP 73, 453-455. These variables subsequently became known as DF And, EE And and UX Psc, see VSNET-UNKNOWN 18-20 for details via: http://vsnet.kusastro.kyoto-u.ac.jp/vsnet/Mail/vsnet-unknown/maillist.html]. All 3 are constant. DZ And : Figure 3 of the above article presents the only two known "dips" of this star, as noted by Goswami et al 1998 [see vsnet-unknown 18]. They occur in isolation as negative plate detections, and are _prior_ to all the positive plate detections. Further, Goswami et al 1998 find no spectral evidence of a type RCB star, the suggested type. AFOEV archival data [via CDS, Strasbourg] from approx. JD 2441500 to 2450500 has 2977 positive observations with a spread of around visual magnitude 9.4 to 10.4, though when plotted this looks merely like the spread from different observers. Indeed, period analysis provides no periodicity over and above a hint of the "annual pseudo-alias" at 365 days. No dip occurs in the data. EE And : Little data exists for this star either at the AFOEV archive or the VSOLJ archive [via VSNET, Kyoto]. Mostly it consists of negative observations. Other than that there appear to be some visual observations around the mag 13.8 - 14 mark, with 3 apparently aberrant 15th mag observations. I am going to assume it is constant. UX Psc : I noted in VSNET-UNKNOWN 19 that UX Psc is really a QSO... ...in fact that is incorrect. I cannot find it in any QSO catalogue. It is in fact simply a Galaxy, and listed in the Lick Northern Proper Motion Galaxies Catalogue... ...their is some eastward error in the GCVS4 position which confused me at first. I _especially_ note that Cragg decided this "star" was a red LPV of 300 day period! In Conclusion : DZ And, EE And and UX Psc probably appeared variable on photographic plates due to insufficient care being taken re colour response. John