21 Ἡ χάρις τοῦ Κυρίου Ἰησοῦ Χριστοῦ* ¦ 1 ℵA,NU ¦ ημων 12 TR μετὰ πάντων τῶν ἁγίων.* ¦ 23 ℵ ¦ 1 A,NU ¦ 1 υμων TR Ἀμήν.*35 Ma,b,dpt,e⅔,f,g,h,i ℵ,CP,RP,HF,OC,TR ¦ --- Mdpt,e⅓ A,NU,* (London: Bernard Quaritch, Ltd., 1929; 2 vols.). I follow H-F in using Ma-e. Mf-i are my invention—Mf was included by Hodges in his Ma and he ignored Mg-i (H-F is also based on Hoskier). For the composition of the M families see the Appendix. Many minor variants have not been included in the apparatus, which is why the fractions for a family do not always add up to 100%. Where a fraction is ¼ or less I show the family without fraction for the other variant. The citation of f35 is based on the 32 MSS that Hoskier assigned to Mc (his "Complutensian") [but he did not cite 243, and 231 is a mere fragment containing only the first two verses, but has the diagnostic reading at the end of verse 2]. To these the following ten MSS should be added: 1064, 1903 [his 243], 2201, 2323, 2431, 2434, 2554, 2656, 2669 and 2723. After determining that they belonged to the family, by checking a list of diagnostic readings, I then checked them individually at the ±110 places where three or more family members share a variant. At the 40 or so places where the dissent involved seven or more MSS I re-checked most of the MSS done by Hoskier [I had to make a number of corrections, but nothing serious]. I have taken account of every extant MS that belongs to Family 35, so far as I know. Please see the last footnote for Matthew for further information.