_Skull._--Male (based on 5 adults from eastern Greenland): See measurements. As described in _Mustela erminea richardsonii_ except that: Weight more (not recorded); basilar length, 41.3 (39.0-42.4); length of tooth-rows more than length of tympanic bulla; breadth of rostrum measured across lacrimal processes averaging more than a third of basilar length; interorbital breadth more than distance between glenoid fossa and posterior border of external auditory meatus; zygomatic breadth more than distance between last upper molar and jugular foramen. Female (based on 2 adults, Turner Sund and Kap Hoeg): See measurements. As described in _Mustela erminea arctica_ except that basilar length 36.8 (35.9, 37.8), and length of tooth-rows not more than length of tympanic bulla. Skulls of females not in hand when this comparison is written; only the recorded measurements are available. To me the skull of _polaris_ is indistinguishable from that of _arctica_. Therefore the comparisons made of the skull of _arctica_ with those of other subspecies will apply also for _polaris_. _Remarks._--In view of the heretofore erroneous assignment of the type locality of _Mustela erminea audax_ to Greenland, pains were taken to verify the statement by Barrett-Hamilton (1904:393) relative to the type specimen of _polaris_. Taking pains thus seemed the more worthwhile because in the specimen register at the British Museum of Natural History, there is written to the right of catalogue numbers 78-6 = 19 nos. 1-11, "Discovery Bay Presented by Mr. Hart Arctic Collection." This refers to no. 78.6.19.1. There are no ditto marks below but by implication this data applies also to nos. 1-11, which include the holotype of _polaris_. A label attached to the specimen does however give the locality as "Hall Land" "N Greenland" and another label has on it "Ermine, procured by Mr. Beaumont Greenland Lat 89° Long W 59-20." The 89° is obviously a mistake (on the label or in my transcription of it) for 82°.