v 7 f. to pneuma kai to udor kai to aima] in terra, spiritus [et] aqua et sanguis, et hi tres unum sunt in Christo Jesu: et tres sunt qui testimonium dicunt in caelo, Pater Verbum et Spiritus m tol cav; also omitting in Christo Jesus, and reading sicut [et] tres, various MSS of vg. lat, with slight variations, as dant for dicunt. In q, which has not lost nearly half of each line, unum...tres seems to have dropped out by homoeoteleuton, leaving the presence or absence of in Christo Jesu uncertain; the only other differences from m are et aqua and (with Cassiod Epiph. Cant) testificantur. The later MSS of lat. vg. transpose the clauses, reading in caelo, Pater Verbum et Spiritus Sanctus, et hi tres unum sunt: et tres sunt qui testimonium dant in terra, spiritus et aqua et sanguis, many of them omitting the clause which ends v. 8, et hi tres unum sunt. Two late Greek cursives contain the interpolation in forms which are manifestly translations from this latest state of the Latin Vulgate, 162 (about Cent. XV), a Graeco-Latin MS, and 34 (Cent. XVI). In fulfilment of a rashly given pledge, Erasmus introduced it into the text of his third edition on the authority of 34, keeping however the genuine kai oi treis eis to en eisin at the end of v. 8. Various crudities of language were subsequently corrected, partly by the help of the Complutensian text, which was a third independent rendering of the Latin Vulgate into Greek; till at length, by editorial retouching without manuscript authority, the interpolation assumed the form which it bears in the 'Received Text', en to ourano, o Pater, o Logos, kai to Agion Pneuma, kai outoi oi treis en eisi kai treis eisin oi marturountes en te ge, followed by to pneuma kai to udor kai to aima.

There is no evidence for the inserted words in Greek, or in any language but Latin, before Cent. XIV, when they appear in a Greek work written in defence of the Roman communion, with clear marks of translation from the Vulgate. For at least the first four centuries and a half Latin evidence is equally wanting. Tert. and Cyp. use language which renders it morally certain that they would have quoted these words had they known them; Cyp. going so far as to assume a reference to the Trinity in the conclusion of v. 8 (et iterum de Patre et Filio et Spiritu Sancto scriptum est Et tres unum sunt), as he elsewhere finds Sacramenta Trinitatis in other occurrences of the number three (Dom. Orat. 34), and being followed in his interpretation more explicitly by Aug. Facundus, and others. But the evidence of Cent. III is not exclusively negative, for the treatise on Rebaptism contemporary with Cyp. quotes the whole passage simply thus (15: cf. 19), quia tres testimonium perhibent, spiritus et aqua et sanguis, et isti tres unum sunt. The silence of the controversial writings of Lucif. Hil. Amb. Hier. Aug. and others carries forward the adverse testimony of the Old Latin through the fourth into fifth century; and in 449, shortly before the Council of Chalcedon, Leo supplies positive evidence to the same effect for the Roman text by quoting vv. 4---8 without the inserted words in his epistle to Flavianus (Ep. xxviii 5). They are absent from lat. vg. according to its oldest MSS am fu and many others, as also from from the (Vulgate) text of the gallican (Luxeuil) Lectionary.

The words first occur at earliest in the latter part of Cent. V, that is, about the time of the persectuion in N. Africa by the Arians Vandals. They are quoted in part in two of the works attributed on slender grounds to Vigilius of Thapsus (one of which has the whole passage, with the curious variations in terra, aqua sanguis et caro, et tres in nobis sunt), and in an argumentative libellus found in the MSS of the History of Victor of Vita (written about 484), and professing to be a memorial presented in 483,but now justly suspected of being a different work, inserted afterwards (Halm p. 26, referring also to Papencordt). The conventional date of this obscure and as yet unsifted group of controversial writings rests on little evidence, but it is probably not far from the truth. At all events a quotation of some of the disputed words occurs early in Cent. VI in another North African work, written by Fulgentius of Ruspe; and soon after the middle of Cent. VI they stand paraphrased in the Complexiones of cassiodorius, written in the southern extremity of Italy. A prologue to the Catholic Epistles, falsely professing to be written by Jerome,impugns the fidelity of Latin translators, accusing them especially of having placed in their text the 'three words' aquae sanguinis et spiritus only, and omitted Patris et Filii et Spiritus testimonium. This extraordinary production is found in the Fulda MS written at Capua in 546,7 (E. Ranke in his ed. p. viii), the biblical text of which is free from the interpolation, as well as in many later MSS, and probably belongs to the Vigilian period and literature. Even after Cent. VI the references to the inserted words are few till Cent. XI.

The two Old Latin MSS in which they are extant have texts of a distinctly late type: they are q, of Cent. VI or VII (Ziegler) and m, of Cent. VIII or IX (Tregelles, Reifferscheid, Hartel), m being in strictness only an arranged collection of quotations from an Old Latin MS. A MS like that which supplied m with its text must have contributed the foreign element to the common ancestor of the Toledo and La Cava Vulgate MSS; and it is remarkable that m quotes the spurious Ep. of St. Paul to the Laodicenes, which is included in both these copies of the Vulgate.

These two interesting MSS likewise illustrate the manner in which the interpolation probably arose. After v. 9 tol adds these words, quem misit salvatorem super terram, et Filius testimonium perhibuit in terra scripturas perficiens: et nos testimonium perhibemus quoniant vidimus eum, et annuntianus vobis ut credatis; et ideo qui &c.: and in v. 20 after venit they both add ( with m, two London MSS cited by Bentley, and virtually Hill) et carnem induit nostri causa, et passus est, et resurrexit a mortuis, adsumpsit nos, et dedit &c. Paraphrastic interpolations like these argue strange laxity of transcription, such as we find elsewhere in the quotations from the Catholic Epistles in m; but they do not imply deliberate bad faith; and the interpolation of verses 7 and 8 doubtless seemed to its author merely to place explicitly before future readers an interpretation which he honestly supposed to give the true sense of the passage, as it had been indicated by Cyprian and expounded by Cyprian's successors. This interpretation was the more plausible since the Latin text did not contain the significant eis of the original (omitted likewise by Cyp. al and apparently others), which probably was early lost after treis; and it is no wonder that controversial associations should lead Latin readers to assume such words as et tres unum sunt to contain a reference to the Trinity. Even in Greek there are traces of a similar interpretation: one scholiast writes eis Theos, mia Theotes in the margin of v. 8; and another first explains the spirit, water, and blood, and then adds OI TREIS de eipen arsenikos oti sumbola tauta tes triados k.t.l.

The adverse testimony of Greek MSS and of all the oriental versions is supported by the silence of all the Greek fathers; and positive evidence is added by Cyr. al, who three times transcribes verses 7 and 8 with the context (Thes. 363; Fid. 95; Nest. 143).

The most essential facts as to the history of the reading were well set forth by Simon in 1689 (Hist. Crit. du Texte du N.T. 203 ff.). The evidence as enlarged by Mill and Wetstein was rigorously examined by Porson (Letters to Travis) in 1790; and admirable expounded afresh in a more judicial spirit by Griesbach in his second edition (ii App. I --- 25) in 1806. Three new and interesting testimonies on behalf of the inserted words have subsequently come to light, those of m in 1832, of q in 1875, and of the occurrence of the Pseudo-Hieronymic Prologue in fu in 1868. They all however unaffected the limit of date which was indicated by Simon and fixed by Porson.