Matt 7:21 ουρανοις

Some early manuscripts (ℵ B C Z Δ Θ 0281vid f1 33. 892. 1424) and most editors (Bover, Greeven, Lachmann, Merk, Soden, Tischendorf [8th], Vogels) naturally add τοις before ουρανοις, but authorial usage indicates that after this particular construction with the genitive, the anathrous ουρανοις (without the article) is to be preferred, as evidenced by most pre-10th-century witnesses (E G K L M O S U V W X Δ Π Σ Φ Ω 047 Byz f13.35 565. 1500. 2224) and Tischendorf (7th). In considering Matthew's style for the construction πατηρ + PRONOUN + ARTICLE + εν +/- ARTICLE + ουρανοις, only the indisputable passages are considered (i.e., where the Nestle-Aland and Robinson-Pierpont editions agree), laying aside temporarily the four disputable passages (i.e., 5:45; 7:21; 10:32, 33). What is discovered is that the article always precedes ουρανοις except when πατηρ and its accompanying articles are in the genitive case:

nominative/vocative:
ο πατηρ υμων ο εν τοις ουρανοις/ο ουρανιος (5:48; 7:11)
Πατερ ημων ο εν τοις ουρανοις (6:9)
ο πατηρ μου ο εν τοις ουρανοις (16:17)
[ο πατηρ υμων/μου ο ουρανιος (6:14, 26, 32; 15:13; 18:35)]

dative:
τω πατρι υμων τω εν τοις ουρανοις (6:1)
τω πατρι σου τω εν τω κρυπτω (6:6, 18)

accusative:
τον πατερα υμων τον εν τοις ουρανοις (5:16)

BUT:

genitive:
του πατρος μου του εν ουρανοις (12:50; 18:10, 19)
του πατρος υμων του εν ουρανοις (18:14)

     Allowing Matthew's indisputable custom to inform the disputable cases leads to the probability that in those three passages where after the genitive construction (i.e., after του πατρος ... του εν ...) a minority of witnesses adds the article before ουρανοις (i.e. Matt 7:21; 10:32, 33), the addition should be considered secondary. Likewise in Matt 5:45 τοις, where a divided majority of witnesses adds the article after the same construction, the addition of τοις is likely secondary. In all four of these variations, the addition of the article may be seen as assimilation to the pattern following the constructions with all of the other cases, i.e., after the nominative, dative, accusative, and vocative constructions. Moreover, preference for assimilation as the likeliest cause of the addition of τοις in the present case is increased due to the presence of τοις in every preceding such construction in Matthew up to this point (5:16, 48; 6:1, 6, 9, 18; 7:11), with the nearest such occurrence just 10 verses earlier (7:11). See the note on Matt 5:45 τοις for an initial discussion of the textual problem.